Dissertation
submitted to the
Combined Faculties for the Natural Sciences and for Mathematics
of the Ruperto‐Carola University of Heidelberg, Germany
for the degree of
Doctor of Natural Sciences
Presented by
Master of Sciences Philipp Konstantin Zimmermann
Born in Berlin‐Zehlendorf
Oral examination: April 16th, 2015
Genome‐wide detection of induced DNA double strand breaks
Referees:
Prof. Dr. Christof von Kalle
Prof. Dr. Stefan Wiemann
I
1. INTRODUCTION 1
1.1 The DNA damage response and genomic instability 1 1.1.1 DNA damaging agents and the DNA damage response 1
1.2 DNA double‐strand break (DSB) repair pathways 3 1.2.1 The Non‐Homologous End Joining (NHEJ) Repair Pathway 4 1.2.2 Methods for the detection of DNA damage and DNA repair activity 6
1.3 Radio‐ and chemotherapy 7 1.3.1 Types of ionizing radiation 7 1.3.2 Radiation‐induced DNA damages 8 1.3.3 The topoisomerase family 9 1.3.4 Topoisomerase 2 catalytic cycle and targeting by anticancer drugs 10 1.3.5 Cancer therapy‐induced delayed genomic instability 11
1.4 Lentiviruses 11 1.4.1 History and phylogeny of lentiviruses 11 1.4.2 The lentiviral genome 12 1.4.3 The life cycle of lentiviruses 13 1.4.4 Structure of lentiviral vectors for gene therapy 15
1.5 Scientific aims 16
2. MATERIALS AND METHODS 18
2.1 Materials 18 2.1.1 Chemicals 18 2.1.2 Enzymes 19 2.1.3 Bacteria 19 2.1.4 Cell lines and primary cells 19 2.1.5 Antibodies 20 2.1.6 Plasmids 20 2.1.7 Oligonucleotides 20 2.1.7.1 Standard primers for q‐RT‐PCR 20 2.1.7.2 Primers used for linker cassettes in LAM‐PCR 21 2.1.7.3 Primers used for LAM‐PCR 21 2.1.7.4 Fusionprimer for Pyrosequencing 21 2.1.7.5 Primers and oligos used for direct DSB labeling approaches 22 2.1.8 Commercial kits 22 2.1.9 Buffers, Media, Solutions 23 2.1.10 Disposables 24 2.1.11 Equipment 24 2.1.12 Software and data bases 25
2.2 Methods 26
II
2.2.1 Cell Culture Methods 26 2.2.1.1 Cell Cultivation 26 2.2.1.2 Freezing and thawing of cells 26 2.2.1.3 Cell Counting 26 2.2.1.4 Transfection 27 2.2.1.5 Transduction 27 2.2.1.6 Virus production 28 2.2.1.7 Determining the lentiviral titer on Hela cells 28 2.2.1.8 MTT Assay 28 2.2.1.9 Immunostaining of H2AX foci 29 2.2.1.10 Inhibition of NHEJ‐repair activity by Nu7441 and Mirin 29 2.2.1.11 Irradiation of cells 30 2.2.1.12 Inhibition of Topoisomerase 2 in mammalian cells with doxorubicin and etoposide 30 2.2.1.13 Preparation for FACS 30 2.2.1.14 Synchronization of NHDF‐A in G1/G0, S and G2/M phase of the cell cycle 31 2.2.2 Molecular Biology Methods 32 2.2.2.1 Isolation of genomic DNA from cultivated cells 32 2.2.2.2 Determining the DNA concentration 32 2.2.2.3 Polymerase Chain Reaction (PCR) 32 2.2.2.4 DNA agarose electrophoresis 33 2.2.2.5 DNA isolation from agarose gel 33 2.2.2.6 Absolute quantitative real‐time PCR (q‐RT‐PCR) 33 2.2.2.7 Linear‐Amplification Mediated Polymerase Chain Reaction (LAM‐PCR) 34 2.2.2.8 Cleaning‐Up of PCR products using AMPure XP beads 39 2.2.2.9 Fusionprimer‐PCR 39 2.2.2.10 SureSelect Target Enrichment for Illumina Multiplexed Sequencing 40 2.2.2.11 Tdt‐mediated labeling of DSB sites 44 2.2.2.12 Biotin Quantification 47 2.2.2.13 Linker‐Amplification‐Mediated DSB Trapping (LAM‐DST) 48 2.2.2.14 Pyrophosphate sequencing 50 2.2.2.15 Cloning of PCR amplicons using the TOPO‐TA Cloning Kit 50 2.2.2.16 Transformation of circular DNA into chemically‐competent E.coli 50 2.2.2.17 Mini‐ and maxipreparation of plasmid DNA 51 2.2.2.18 Enzymatic DNA restriction digest 51 2.2.3 Bioinformatical Methods 51 2.2.3.1 Automated Sequence Analysis (HISAP) 51 2.2.3.2 A549 mRNA expression analysis 52 2.2.3.3 DNaseI Hypersensitive Sites 52 2.2.3.4 Histone modifications and Transcription Factor Binding Sites 52 2.2.3.5 Ingenuity Pathway Analysis (IPA) 53 2.2.3.6 Identification of DSB site clusters in the genome 53
3. RESULTS 54
3.1 IDLV‐mediated capturing of radiation‐induced DSB sites in vivo 54
III
3.1.1 Immunostaining of H2AX foci in irradiated cells 54 3.1.2 MTT assay to determine lethal dose values for etoposide and doxorubicin 55 3.1.3 IDLV‐Delivered DNA‐baits tag radiation‐induced and repaired DSB sites 56 3.1.4 IDLV DSB trapping in NHDF‐A with impaired NHEJ‐repair activity 58 3.1.5 Trapping and mapping of TOP2 poison‐induced DSB 58 3.1.6 Calculating the number of integrated IDLV copies per irradiated cell 59 3.1.7 SureSelect Target Enrichment for analyzing IDLV vector integrity 60
3.2 Analyzing early DSB repair events and kinetics of DSB induction at single nucleotide resolution 61 3.2.1 Genomic tagging of early‐repaired radiation‐induced DSB sites by IDLV 62 3.2.2 Synchronization of NHDF‐A and Hela cells in G1, S and G2 phase 62 3.2.3 Tdt‐mediated labeling of radiation‐induced DSB sites 63 3.2.4 Linker‐Amplification‐Mediated DSB‐Trapping for DSB labeling in real‐time 66
3.3 DSB site distribution in radiation‐surviving and expandable cell populations 68 3.3.1 Identification of radiation‐induced DSB sites by LAM‐PCR 68 3.3.2 DSB are not enriched on chromosomes, in genes and gene‐regulatory regions 69 3.3.3 IDLV integration at radiation‐induced DSB sites is mediated by NHEJ‐repair 71
3.4 Analyzing the influence of transcriptional activity, chromatin status and gene classes and networks on DSB site distribution 72 3.4.1 Trapping of Radiation‐Induced DSB is not influenced by transcriptional activity 72 3.4.2 The location of radiation‐induced DSB sites is composed of histone modifications defining active chromatin 73 3.4.3 Radiation‐induced DSB in radiation‐survivor cells are enriched in genes and networks regulating cell survival 77 3.4.4 Identification and analysis of radiation‐induced DSB sites over time 79
3.5 Identification of frequently damaged and repaired genomic regions 81 3.5.1 DSB Trapping in irradiated and passaged cells reveals common regions of radiation‐induced and repaired damage 81 3.5.2 Radiation‐related DSB hotspots overlap with genes involved in maintaining genome stability and DNA repair 84 3.5.3 DSB hotspots overlap with eu‐ to heterochromatin border regions 86
4. DISCUSSION 88
4.1 Immunostaining of H2AX is not suitable to detect DSB and genomic instabilities in cancer therapy surviving cell populations 88
4.2 NHEJ‐mediated IDLV integration at DSB sites stably marks DNA damage and repair sites in living cells 89
4.3 Identification of radiation‐induced DSB sites 90 4.3.1 Transcriptional activity before irradiation does not influence DSB site distribution 90 4.3.2 DSB site distribution is non‐random with respect to the genome accessibility 91 4.3.3 Genes involved in specific cellular processes are enriched for induced DSB 93
IV
4.3.4 Clonality of DSB site distribution over time 93 4.3.5 Radiation‐induced and repaired DSB sites cluster in hotspots in specific genomic regions 94
4.4 Methods for in situ labeling of induced DSB sites 96
5. SUPPLEMENT 98
5.1 Supplementary Figures 98
5.2 SupplementaryTables 110
5.3 Figure Index 135 5.3.1 Figures 135 5.3.2 Supplementary Figures 136
5.4 Table Index 137 5.4.1 Tables 137 5.4.2 Supplementary Tables 137
5.5 Zusammenfassung 139
5.6 Summary 140
5.7 Abbreviations 141
5.8 References 146
5.9 Publications and congress attendances 152
6. DANKSAGUNG 154
1 INTRODUCTION
1. INTRODUCTION
1.1 The DNA damage response and genomic instability
1.1.1 DNA damaging agents and the DNA damage response
The DNA in the cell encodes the genetic information required for the functioning of all cells and living organisms. Every day, the DNA is under constant pressure to be destabilized by various processes and agents. These lesions can block DNA replication and transcription and are broadly divided according to their origin into either endogenous or exogenous (Figure 1) [1]. Endogenous processes such as DNA replication by DNA polymerases induce low levels of DNA damage. Other endogenous sources of DNA damage include genomic fragile sites, nucleases and reactive oxygen species (ROS) such as superoxide and hydrogen peroxide stemming from metabolic processes in the mitochondria [2]. These latter, highly reactive molecules form chemical bonds with other molecules within seconds of their production. Among their target molecules, the DNA is a susceptible target, in which ROS can cause base and sugar modifications, DNA‐protein crosslinks, single‐strand and double‐ strand breaks. In addition to endogenous processes, numerous exogenous sources also threaten the genomic integrity (Figure 1). The most pervasive form of exogenous DNA damaging agent is ultraviolet (UV) rays originating from sunlight that can hit the DNA directly and induce up to 100,000 DNA lesions per exposed cell [2]. These UV light generate various forms of DNA modifications, the most stable being the covalent link between two pyrimidine DNA bases [2]. X‐rays represent another type of DNA damaging agent, which is often used for imaging and cancer radiation‐therapy. These rays generate radicals in the DNA or in other cellular molecules that attack DNA bases, causing DNA oxidations as well as single‐strand and double‐strand breaks. Additional examples of exogenous DNA damaging agents include benzo‐α‐pyrene, polycyclic‐aromatic hydrocarbons and cancer chemotherapeutics that interfere with the DNA replication or the DNA damage repair machinery.
Along with the various types of DNA damages induced by endogenous processes and environmental agents, DNA double strand breaks (DSB) are the most dangerous DNA lesion [1]. In contrast to other types of DNA damage, DSB directly threaten genomic integrity, because they disrupt the continuity of the DNA. Failure to efficiently repair such DNA lesions can compromise genome integrity and result in a variety of genomic changes. Amongst others these include point mutations, chromosome number variations, gross chromosomal rearrangements such as translocations, amplifications, insertions, deletions, dimeric oncogenes and copy‐number variations, which can initiate neoplastic transformation [1]. In order to maintain genomic integrity and counteract the damage inflicted on the genome, the cell triggers a coordinated network of signaling cascades, which are collectively termed the DNA damage response (DDR) (Figure 1). 2 INTRODUCTION
Endogenous Exogenous Agents DNA Damaging Processes Agent Alkylating ROS Irradiation UV Agents
DNA Damage
ICL, DNA PD, Oxidation SSB, DSB adducts 6,4-PP
Sensor DNA damage Sensor Proteins
Checkpoint DNA Damage Checkpoints
Chromatin Remodelling Transcription DNA Repair Responses Apoptosis Senescence Cell Cycle Arrest
Figure 1: DNA damaging agents and the DNA damage response in mammalian cells. DNA damage can either arise by endogenous processes such as inflammation, reactive oxygen species, and DNA replication or can be induced by irradiation, UV light, or chemicals. The DNA damage inflicted is diverse, ranging from oxidized DNA bases to interstrand cross links (ICL), DNA adducts, single‐strand (SSB) and double‐strand breaks (DSB), pyrimidine dimers (PD) and 6,4‐photoproducts (PP). These damages activate DNA damage sensors that in turn activate various cellular responses which are collectively termed the DNA damage response (DDR). In addition to pathways dedicated to removing DNA modifications by DNA repair, cells have evolved several pathways to activate cell cycle checkpoints, block or induce transcription and remodel local chromatin structures. Moreover, if the DNA damage is too severe, apoptosis is induced. Image modified from [3].
The DDR can be divided into different sequential steps. First, the DNA damage sensor proteins of the DDR sense DNA lesions and bind to the damaged site. Subsequently, these proteins transmit the signal of the DNA damage to effector proteins by modifying these proteins post‐translationally. The phosphorylation, acetylation, ubiquitination and sumolyation of the effector proteins leads to an amplification of the damage signal. This signalling blocks cell cycle progression and allows the cell sufficient time to repair the damage. Additionally, the transcription of DDR‐related proteins can be activated and the transcription of non‐DDR‐related proteins as well as replication blocked. Furthermore, the DDR promotes alterations in local chromatin structure by post‐ translationally modifying the histone components in order to reveal the underlying DNA sequence and to recruit the DNA repair machinery (Figure 1). Histone modifications include phosphorylation, acetylation, methylation, and ubiquitination [4]. Cells that are defective in any of the aforementioned response mechanisms are more sensitive to DNA damaging agents, and many of such defects cause cancer, neurological defects and pre‐mature aging [1]. The first observation that inefficient DNA repair may increase the probability for cancer initiation was documented in 1968 by James Cleaver who reported that patients suffering from Xeroderma Pigmentosum (XP) have a defect in the DNA repair pathway responsible for repairing UV light‐induced DNA damages [5, 6]. Since this discovery, several cancer types stemming from mutations in the DNA damage response have been identified (Table 1). 3 INTRODUCTION
Table 1: Human disorders associated with defects in genome maintenance and enhanced cancer susceptibility. BER: base‐ excision repair; ICL: interstrand crosslink; MMR: mismatch repair; RECQ: recombination Q. Table taken and modified from [3] Disease Abbreviation Mutated Gene(s) Impaired Pathway Ataxia telangiectasia AT ATM DSB repair Atypical Werner WS WRN nuclear structure syndrome Bloom’s syndrome BLS BLM RECQ helicase Dyskeratosis DKC DKC1, TERC1 telomere maintenance congenita FANCA, B, C, D1 (BRCA2), D2, E, F, G, Fanconi anemia FA I, J (BRIP1), L, M, N (PALB2), O ICL repair (RAD51C), and P (SLX4) Li–Fraumeni many (p53 tumor p53 Syndrome suppressor inactivation) Nijmegen breakage NBS NBS1 DSB repair syndrome Rothmund–Thomson RTS RECQL4 RECQ helicase syndrome RECQ helicase, influencing Werner syndrome WS WRN nuclear structure, DSB repair, ICL repair, MMR, BER Xeroderma XP XPA‐G NER pigmentosum
The plethora of DNA damages induced by endogenous processes and exogenous agents necessitates the coordinated action of multiple distinct DNA repair pathways. Some DNA lesions can be directly reversed; most DNA damages however require the activity of several repair proteins and processing of the damaged DNA [2]. During DNA mismatch repair, a single strand break is induced upon DNA detection and the single strand removed before polymerase and ligase enzymes fill up the gap and reseal the DNA. In base‐excision repair, DNA glycosylase enzymes recognize the damaged DNA base, and initiate base removal and repair by nuclease, polymerase and ligase proteins. The nucleotide excision repair (NER) system recognizes helix‐distorting base lesions, excises a 22‐30 base oligonucleotide and thereby produces a single‐stranded DNA that is subsequently repaired. Upon repair of the DNA damage, the DDR proteins dissociate from the damaged and repaired siteallowing the cell to re‐enter the cell cycle [2]. If the DNA damage is too severe and cannot be efficiently repaired, cells can either engage in tolerance pathways which permit survival at the cost of mutagenesis [7] or induce cell death by apoptosis [1].
1.2 DNA double‐strand break (DSB) repair pathways
For the repair of DSB, two different mechanisms are present in the cell: Non‐Homologous End Joining (NHEJ) and Homologous Recombination (HR), which can be both subdivided into several pathways. The HR and NHEJ pathway differ mechanistically in several ways. HR repair is initiated by the generation of a single‐stranded DNA (ssDNA) overhang, promoted by the Mre11‐Rad50‐Nbs1 (MRN) complex and CtBP‐interacting protein (CtIP), whereas NHEJ‐repair is activated by binding of the Ku70/80 heterodimer to the DSB site. Moreover, NHEJ‐repair is considered error‐prone, because free DNA ends are quickly rejoined without requiring homology, whereas in HR the homologous sister chromatid is used as a template for copying the genomic information to the damaged DNA strand. Hence, in contrast to NHEJ, HR is not active throughout the cell cycle, but is restricted to late S and G2 phase when the homologous DNA template is available after replication. Thus, NHEJ is the dominant pathway 4 INTRODUCTION
for DSB repair in mammalian cells [2]. The exact cellular mechanisms determining which of the DNA repair pathway is used to repair a DSB are still not fully understood. However, recent work suggests that the choice between classical NHEJ and HR in replicating cells is regulated by DNA end resection [8]. Two regulators of the DSB repair pathway choice have emerged, namely the tumor suppressor proteins 53BP1 and BRCA1. 53BP1 contributes to NHEJ by interacting with chromatin at the DSB sites, thereby inhibiting DNA end resection and tethering the free DNA ends of the DSB in close proximity to promote their ligation [9]. Moreover, 53BP1 binds dimethylated histone H4 (H4K20me2), thereby stabilizing a chromatin conformation that is non‐permissive to nuclease access and limiting DNA end resection of the DSB intermediates [10]. Since in S/G2 phase the level of HR increases, mechanisms exist that block or reduce the activity of 53BP1 at DSB sites in order to promote HR. A key protein in this process is BRCA1; however, the exact mechanisms of BRCA1‐mediated 53BP1 inhibition and activation of HR at DSB sites are still unknown. It has been shown that BRCA1 stabilizes CtIP, a key molecule required for DSB end resection that promotes HR [11, 12]. Furthermore, BRCA1 may also influence the local abundance of H4K20me2, which prevents 53BP1 enrichment and NHEJ‐repair [13]. Other results suggest that BRCA1 also increases the level of ubiquitylation at DSB sites which may contribute to a chromatin status promoting HR by granting access of the resection machinery [9].
The DNA repair activity and repair kinetics vary in the different compartments of the genome [14]. Densely compacted heterochromatin (HC) was shown to be refractory to DNA repair factors such as γH2AX. Furthermore, DSB sites in heterochromatin, gene poor or pericentromeric regions in the genome are repaired with slower kinetics than DSB in euchromatin [15]. Thus, it has been hypothesized that the slower DSB repair kinetics in HC are due to hindered access of DNA repair factors to these dense chromatin regions [14]. However, Jakob and colleagues demonstrated that H2AX is phosphorylated inside HC, but that the damaged site is subsequently relocated to regions of lower chromatin density to enable efficient DSB repair [16]. Another factor influencing the kinetics of DSB repair is the chemical DNA damage complexity [17]. Naturally‐occurring DNA lesions tend to be isolated and homogeneously distributed and can thus be repaired efficiently by the DNA repair pathways. In contrast, clustered DNA damages such as those induced by ion particle irradiation have a reduced reparability compared to that of individual lesions [18].
1.2.1 The Non‐Homologous End Joining (NHEJ) Repair Pathway
Non‐Homologous End Joining (NHEJ) is the primary DSB repair pathway in the cell and rejoins DNA ends regardless of their homology, which increases the probability for genomic aberrations such as deletions, insertions and translocations [19]. Thus, NHEJ‐repair is an important DNA repair pathway which contributes to both genome protection and mutations. Classical NHEJ‐repair in mammalian cells was discovered in the 1980s when Mimori and colleagues first described the DNA end binding protein Ku [20]. However, it was not until 1994 when Stamato and colleagues revealed Ku’s function in the DDR by showing that cells lacking Ku are sensitive to irradiation [21]. Moreover, several groups observed that the repair of DSB in NHEJ‐deficient cells resulted in increased deletion frequencies and use of microhomologies at the repair junctions, which points to an additional NHEJ‐like DNA repair mechanism. This second pathway was called microhomology‐mediated end joining (MMEJ) or alternative NHEJ (aNHEJ), and occurs at approximately 10% of the frequency of normal or classical NHEJ (cNHEJ). Furthermore, it is less faithful than cNHEJ, since excessive deletions and chromosomal translocations are frequently found at the repaired DSB sites, with some of them leading to oncogenic transformation [22]. 5 INTRODUCTION
Both classical and alternative NHEJ‐repair can be divided into four basic steps: (1) DSB recognition, (2) end binding and synapsis formation, (3) end processing, and (4) ligation [22] (Figure 2). Classical NHEJ is initiated by binding of the Ku70/80 heterodimer to free DNA ends at the DSB site. Upon recognition of the DSB, the Ku heterodimer translocates along the DNA from the DSB site in order to allow additional Ku dimers to bind to the DNA, which protects the DNA termini from end resection and acts as a scaffold for the recruitment of additional repair factors including ATM, DNA‐PKcs and DNA polymerases. The Ser/Thr protein kinase ataxia‐telangiectasia mutated (ATM) is a key regulator of the DNA damage response and activates a plethora of downstream effectors by phosphorylation. Among its target molecules, 53BP1, H2AX, p53 and DNA‐PKcs are a few of the most intensively studied repair factors [23]. Another kinase, namely DNA‐PKcs (DNA‐dependent protein kinase, catalytic subunit) belongs to the same family as ATM and also phosphorylates several proteins such as H2AX, Artemis and XRCC4. Since these proteins are involved in downstream processes of DNA repair, their activation multiplies the DNA damage signal. In addition to activation of effector molecules for DNA repair, the DNA‐PK‐ Ku70/80 complex undergoes autophosphorylation, resulting in a conformational change that increases the accessibility of the DNA for additional DNA processing enzymes and ligases. Aligned and compatible DNA ends can be directly rejoined by the DNA ligase IV. However, complex DNA ends such as those produced by ionizing radiation require processing by additional enzymes to prepare the DNA termini for ligation. One of these processing enzymes is Artemis, which is directly activated by DNA‐PKcs. In complex with DNA‐PKcs, Artemis removes single‐stranded DNA (ssDNA) overhangs that contain damaged nucleotides. Then, XRCC4 interaction with the polynucleotide kinase/phosphatase (PNKP) results in the phosphorylation of 5’ OH DNA ends and the removal of the phosphate molecule from the 3’ DNA terminus, thereby forming compatible DNA ends for ligation. Subsequently, DNA polymerases fill in the gaps at the DSB site. In the final step of cNHEJ, the two DNA ends are rejoined by the XRCC4‐ligase IV‐XLF complex. XRCC4 stabilizes ligase IV and stimulates its ligation activity. XLF stimulates the ligation of non‐cohesive DNA ends, and is essential for gap filling by the DNA polymerases, suggesting that it plays an important role in the alignment of DNA ends and thus maintaining stability of the broken DNA ends. Upon repair of the DSB, the repair proteins dissociate or are removed from the DNA [22].
In contrast to cNHEJ, the mechanisms and proteins involved in alternative NHEJ remain poorly understood. A characteristic feature of aNHEJ is the use of a 0‐10bp microhomologous sequence during the alignment of the broken DNA ends. Furthermore, distinctive DNA repair signatures found at DSB sites repaired by aNHEJ imply that enzymes involved in aNHEJ promote end resection and include nucleases and ligases. Several proteins have been identified to be involved in end resection (Mre11, CtIP) and ligation (PARP1, DNA ligase III) [22] (Figure 2). 6 INTRODUCTION
Figure 2: Classical and alternative NHEJ‐repair in mammalian cells. During classical NHEJ‐repair, DSB are recognized by the Ku70/80 heterodimer, which recruits other repair factors. DNA‐PKcs multiplies the damage signal by phosphorylating several downstream targets such as Artemis, which processes the DNA ends for subsequent ligation by the XRCC4‐Ligase IV‐XLF complex. In contrast to cNHEJ, most repair factors involved in aNHEJ remain unknown. However, PARP1 is known to recognize the DNA break and Mre11 and CtIP are involved in end resection. Image modified from [22].
1.2.2 Methods for the detection of DNA damage and DNA repair activity
The most frequently used method for the detection of DSB and DNA repair activity is immunostaining of DNA repair proteins with fluorescently‐labeled antibodies and subsequent visualization by microscopy. Several DNA repair factors that assemble at the DSB site form large complexes that can be microscopically visualized. Among the most prominent molecules are DNA‐PK, 53BP1, and γH2AX. The histone variant H2AX is incorporated into the nucleosomes at DSB sites and becomes phosphorylated at its serine residue at position 139 (Ser139) by the kinase activity of DNA‐PK, ATM and ATR. Upon phosphorylation, γH2AX forms large foci, termed radiation‐ induced foci (RIF), which can span up to several megabases around the DSB site in order to facilitate DNA repair and recruit other DNA repair factors. When the DNA damage has been repaired, γH2AX is dephosphorylated and disassembles from the DSB site [2].
Another approach often used to study DNA repair sites and genomic aberrations is whole genome sequencing. In 2005, Roche/454 Life Sciences introduced the first commercially‐available next generation sequencing technology [24] that is capable of generating 80–120 Mb of sequence in 200‐ to 300‐bp reads. The process is divided into several sequential steps, starting with the isolation of genomic DNA, followed by a fragmentation step, the ligation of PCR adapters and separation into single DNA strands. The single‐stranded DNA fragments are subsequently captured on magnetic beads under conditions that favor one DNA fragment per bead. The beads are incorporated into oil droplets containing all PCR reagents that allow amplification, thereby producing millions of copies of a unique DNA template. Subsequently, the droplets are broken up, and each bead is deposit in a single picoliter‐sized fiberoptic well for pyrosequencing. Pyrosequencing is a sequencing‐by‐synthesis approach in which pyrophosphate (PPi) is released from the nucleotide by the DNA polymerase during nucleotide incorporation. The ATP sulfurylase converts the free PPi molecules into a substrate for the luciferase, which results in light emission and thereby determines the DNA sequence. Similar to the Roche/454 sequencing 7 INTRODUCTION
platform, Illumina sequencing is also a sequencing‐by‐synthesis approach. However, it does not require a pre‐ amplification step of the DNA template, since these are directly bound to the surface of a flow cell. The Illumina flow cell is densely populated with forward and reverse PCR primer adapters which are complementary to adapters ligated to the DNA templates. The DNA templates bind to the forward and reverse primer adapters of the flow cell and form a bridge that serves as the substrate for amplification. For sequencing, modified nucleotides with reversible terminators are used. This allows a single nucleotide to be incorporated in each sequencing cycle [25]. Thus, Illumina sequencing has a higher accuracy than the 454 platform. Each nucleotide carries its own chemically‐cleavable fluorescent dye at the 3’ OH terminus. The cleavage of the terminator results in fluorescence that is recorded by a camera. Recent advancement of the reversible‐terminator sequencing technology is the MiSeq sequencing by Illumina, which produces 2 x 300 paired‐end reads in a single run, allowing small genome sequencing and assembly. Further, it allows processing of more samples and generates more reads per run than previous Illumina sequencing platform [26].
1.3 Radio‐ and chemotherapy
1.3.1 Types of ionizing radiation
Ionizing radiation (IR) is defined as the radiation that has enough energy to remove electrons from atoms, and thereby ionizes them [27]. The IR energy is released inside the cell and is absorbed by various molecules such as proteins, lipids and the DNA and ionizes them. Radiation can be either directly or indirectly ionizing. During direct ionization, the radiation beam hits the target molecule in the cell and disrupts its structure. In contrast, during indirect ionization, the energy of the radiation beam is transferred to an electron of an atom, which is not part of the target molecule. This electron becomes excited, and, due to its higher energy, is lost from the atom. The ionized atom forms a radical that causes damage to different molecules inside the cell [7, 27] such as the DNA.
In general, two types of IR can be distinguished: photon and hadron. X‐rays and α‐rays belong to the group of photon radiation. X‐rays are generated by rapid stopping of highly accelerated electrons, have a wavelength between 0.01 and 10nm and energies in the range of 102 to 105 electron volt (eV). In contrast, hadron radiation consists of particles including protons, neutrons, and heavy charged ions such as carbon (12C) and ion (56Fe) atoms [7]. The two different types of radiation differ in their biological effectiveness which is expressed by the relative biological effect (RBE) value. The RBE is defined as the dose of a particular radiation divided by the dose of X‐rays required for an equal biological effect [27]. The higher the RBE for a specific radiation type, the more damage is induced per unit of energy deposited in the cell/tissue. In comparison to photons, carbon ions have an increased RBE, which is calculated between 2 and 5 depending on the cell type [28]. The different RBE of photons and hadrons result from the distances travelled by the IR in tissues and the pattern of ionizing events along the track. Photons are only sparsely ionizing, dispose their energy in atoms spaced by several hundred nanometers apart and cannot penetrate the tissue deeply. Thus, ionization events occur most at the tissue surface. In contrast, hadrons leave a dense trail of ionized atoms along their path that are spaced only about a tenth of a nanometer apart [29]. Moreover, hadron particles enter into the tissue deeply and only release small amounts of energy at the beginning of their route. At the end of their path, these particles rapidly decelerate and release high amounts of energy before stopping completely [30]. Since hadrons deposit their energy at a depth proportional to the energy of the charged particle [7], hadron irradiation allows a more precise deposition of energy inside the target tissue than photon irradiation. 8 INTRODUCTION
1.3.2 Radiation‐induced DNA damages
In order to determine the density of ionizations along the radiation track and thus the extent of radiation damage the Linear Energy Transfer (LET) is used as a measure. The LET is defined as the average energy deposition per unit length of the ionizing track, and its unit is keV/µm. Photons have a low LET, whereas in general hadrons have a high LET. Both low and high LET radiation induce a variety of DNA lesions including DSB, and the complexity of the DNA damage increases with increasing LET. Isolated DNA damages mainly occur after low LET irradiation, are spaced at a larger distance from other damages and can be repaired efficiently by the DNA repair machinery [29] (Figure 3).
High LET Low LET Excitation Ionization Radiation track Clustered DNA damage DNA damage
Isolated DNA lesions
Figure 3: Induction of ionizations and DNA damage by high‐ and low LET irradiation. High LET radiation causes more ionizations and excitations at high density than low LET radiation that creates more isolated DNA damages.
In contrast, high LET irradiation also induces complex clustered DNA lesions that include two or more lesions within 20bp or one helical turn (Figure 3). Amongst others, these DNA damages comprise abasic sites (AP), damaged bases and single‐strand breaks (SSB) that can be converted into potentially lethal DSB when the DNA repair is disrupted. Whether a non‐DSB cluster is converted into a complex DSB depends on the type of lesion induced, the distance separating the DNA lesions and whether the additional lesions disrupt the binding of DNA repair enzymes to other DNA damage sites [7]. Thus, in order to reduce the probability of DSB formation during damage repair, opposing DNA lesions are repaired sequentially [31]. Inefficient DNA repair of DNA damage clusters can result in increased levels of mutagenesis. The overexpression of certain DNA repair proteins such as the DNA glycosylase responsible for repair initiation can convert an oxidative DNA damage to a DSB (Figure 4). Non‐DSB clusters consisting of opposing AP sites are converted to DSB sites by the activity of the exonuclease Ape1, which stimulates cNHEJ‐repair or Ku‐independent DNA repair activity (Figure 4). Complex DNA damage clusters consisting of oxidized and abasic DNA molecules can be converted into a complex DSB by the nuclease ApeI, making it highly difficult for the cell to efficiently repair the damage [32]. Such complex clustered DNA damages are mainly induced by high LET radiation, are difficult to repair, have slow repair kinetics, and are thus considered to be more detrimental to the cell than randomly distributed DNA lesion induced by low LET radiation [33], thereby explaining the increase in RBE of high LET. In addition to the formation of complex clustered DNA lesions and radiation‐induced DSB, replication‐induced DSB sites are formed after ionizing radiation [34, 35], when a replication fork meets an unrepaired non‐DSB clustered damage site [35, 36]. 9 INTRODUCTION
Intact DNA
Oxidative damage cluster Abasic damage cluster Complex damage cluster O O A A AAA O
BER ApeI ApeI
O O A
BER DSB Complex DSB
Death NHEJ Death NHEJ Ku-independent (Ku, DNA-PK, (Ku, DNA-PK, DSB Repair XLF, XRCC4, XLF, XRCC4, HR, SSA, aNHEJ Repair efficient Ligase IV, Artemis) Ligase IV, Artemis)
Repair not always Repair efficient/accurate inefficient/mutagenic
Figure 4: Overview about repair of clustered DNA damages and repair outcomes. Two oxidative DNA damages (O) can be repaired efficiently by sequential base excision repair activity. However, replication through this damage can increase the mutation frequency of the base damage. Furthermore, in the presence of DNA glycosylases, two oxidative lesions can be converted into a DSB. Non‐DSB, abasic damage clusters are converted into DSB by the nuclease ApeI. DSB sites can induce cell death, or be repaired by error‐prone DSB repair pathways such as classical and alternative NHEJ. Complex DSB clusters resulting from high LET irradiation are more difficult to repair and can result in increased levels of mutagenesis or cell death. A: abasic site; O: oxidative damage; NHEJ: non‐homologous end joining; BER: base excision repair; HR Homologous Repair, SSA: Single‐Strand Annealing
1.3.3 The topoisomerase family
During transcription and replication, the DNA is locally unwound to allow the transcriptional and replication machinery to proceed. This, however, leads to DNA supercoiling and increased tension in adjacent genomic regions, stalling of the replication and transcription machinery as well as formation of abnormal DNA structures. Topoisomerases are enzymes involved in the unwinding of the DNA during transcription and replication. Upon replication, the replicated DNA strands are interlinked and need to be unlinked before cytokinesis can proceed. Failure to disentangle the two DNA strands can lead to genomic aberrations and induction of cell death. In order to protect the cell from replication‐induced damage, DNA topoisomerases also assist in the segregation of the daughter chromosomes before cell division [37].
The human genome encodes six topoisomerases, which are divided into type I and II. Type I topoisomerase (TOP1) only cleaves a single DNA strand, whereas the type II topoisomerase (TOP2) can induce DNA double strand breaks. Both types of topoisomerases cut the phosphodiester backbone of the DNA by a nucleophilic attack from the catalytic tyrosine at the catalytic center of topoisomerase. This reaction is reversible, and the DNA sequence remains unchanged upon religation. Mammalian cells have two TOP2 isoenzymes, namely α and β, which function as homodimers. The expression of TOP2α is closely linked to the cell cycle and the expression increases two‐ to three‐fold during G2/M phase. TOP2β on the other hand, is always expressed in both cycling and non‐cycling cells. The TOP2 isoenzymes have low sequence selectivity, but they preferentially recognize and bind DNA knots, supercoils and interlinks [37]. 10 INTRODUCTION
1.3.4 Topoisomerase 2 catalytic cycle and targeting by anticancer drugs
The TOP2 catalytic cycle is initiated by binding of the TOP2 homodimers to the gate (G) DNA segment, through which the transported (T) DNA segment is passed. After binding to the G segments, TOP2 binds the T DNA segment and two ATP molecules. Subsequently, the homodimer changes its conformation to a closed clamp form, called the TOP2 cleavable complex. Upon binding of Mg2+, the two tyrosyl residues of the TOP2 homodimers attack the phosphodiester backbone in the opposite DNA strands of the G segment to induce a DSB. In this stage, each of two topoisomerase monomers is covalently linked to the 5′‐terminus of an enzyme‐ generated DSB. Then, the T segment is rapidly passed through the G segment and released from the enzyme. After strand passage, the G segment is religated by TOP2, and ATP hydrolysis converts the closed clamp conformation of the enzyme to the open form to release the G segment [37] (Figure 5).
Doxorubicin Etoposide Doxorubicin
Figure 5: The TOP2 catalytic cycle (1) Upon recognition of a DNA crossover region, TOP2 binds to the DNA (G segment, green). (2) ATP binding leads to a conformational shift to a closed clamp form and binding of the T segment (purple). (3) The G segment is subsequently cleaved and (4) the T strand passes through the G segment. (5) TOP2 religates the G segment and (6) returns to its open conformation after ATP hydrolysis. At low concentrations (<1µM), doxorubicin acts like etoposide and inhibits TOP2 DNA religation, whereas at concentrations above 10µM doxorubicin intercalates into the DNA, thereby blocking TOP2 binding. Figure modified from [38].
Since failure to disentangle two interlinked DNA strands can lead to the induction of cell death, TOP2 enzymes are frequently targeted in cancer chemotherapy. Generally, anti‐cancer drugs targeting TOP2 can be classified according to their mechanisms. Molecules that inhibit the religation of the G segment such as etoposide and doxorubicin are termed TOP2 poisons, whereas TOP2 inhibitors such as genistein and azatoxins enhance the formation of the TOP2 cleavable complex. In addition to stimulating DSB formation, TOP2 poisons and inhibitors also induce SSB, affect nuclear processes such as transcription and replication by intercalating into the DNA and by generating reactive oxygen species (ROS). Doxorubicin belongs to the family of anthracyclines, and, at high concentrations (10μM), intercalates into the DNA, which alters the DNA structure and prevents TOP2 from 11 INTRODUCTION
binding to the DNA. At concentrations below 1µM, doxorubicin acts like etoposide and stabilizes the cleavable complex, thereby prolonging the half‐life of the TOP2‐DNA‐intermediate and increasing the possibility of DSB formation. In order for DSB repair to occur, the covalently attached TOP2 enzyme is removed by cellular end‐ processing enzymes. The DNA fragment linked to the peptide is excised by nucleases such as the MRN complex, CtIP or Artemis, and the DSB is repaired by NHEJ [39‐41].
1.3.5 Cancer therapy‐induced delayed genomic instability
In progenitors of radiation‐surviving cells, genomic changes can arise several generations of the initial DNA insult. This phenomenon is known as delayed radiation‐induced genomic instability, and is characterized by the expression of several radiation‐induced effects including apoptosis, gross chromosomal rearrangements, aneuploidy as well as gene mutations and amplification [31]. The mechanisms that initiate and drive delayed genomic instability in progenitor cells are not fully understood yet. However, experimental evidence suggests that increased levels of oxidative stress contribute to radiation‐induced genomic instability, since cells with mitochondrial dysfunction and cells exposed to hydrogen peroxide initiate delayed instability, and treatment with antioxidants can greatly reduce these effects [42, 43]. Moreover, it also possible that delayed genomic instability is caused by delayed DSB induction and subsequent misrepair by NHEJ. Delayed DSB induction and illegitimate joining of DNA ends can generate dicentric chromosomes which block segregation during mitosis, thereby inducing additional DSB and accumulation of mutations [44]. Since radiation‐induced genomic instability is transmitted through the genome for several generations, it was speculated that the initial ionization event is memorized in the genome [45]. Several studies now suggest that radiation induces alterations in local chromatin structures that can lead to increased levels of replication stress and DSB that can initiate genomic instability in radiation surviving cell populations [44, 45]. Despite these results, the functional consequences of delayed genomic instability on carcinogensis and radiosensitivity have not been described to date [44]. Nonetheless, since the genomic effects of delayed genomic instability are similar to those induced directly by irradiation, it is assumed that these genomic changes initiate or drive carcinogenesis [46‐48]. Delayed induced genomic instability has mostly been observed and studied in radiation surviving cells, but, persistent destabilization of the genome following chemical treatment was also reported. For example, bleomycin and neocarzinostatin were equally efficient in inducing delayed genomic instability in progenitors of chemotherapy‐surviving cells [49]. Moreover, anti‐topoisomerase drugs have been reported to induce structural and numerical chromosome aberrations in surviving cells [50]. Thus, both irradiation and chemotherapy can induce delayed genomic instability, which potentially leads to cancer‐therapy resistance and therapy‐induced carcinogenesis.
1.4 Lentiviruses
1.4.1 History and phylogeny of lentiviruses
In 1911, Peyton Rous showed that cancer could be transferred from chicken suffering from sarcoma to healthy chicken by using an ultra filter extracts [51]. Later, it was discovered that viruses in the extract were the cancer‐ inducing agent, and this virus was named after its discoverer Rous Sarcoma Virus (RSV). In 1970, Howard Temin, Satoshi Mizutani and David Baltimore discovered the enzyme reverse transcriptase in RSV, which enables retroviruses to convert their single‐stranded RNA genome into double‐stranded DNA [52, 53]. In the following years, another characteristic feature of retroviruses, the integration of the proviral genome into the host cell’s 12 INTRODUCTION
genome was discovered [54], which is mediated by the viral integrase enzyme. In the following decade, the first human oncogenic retrovirus (HTLV) [55] and the immunodeficiency causing virus HIV [56] were discovered. These viruses belong to the retroviridae family which is divided into two subfamilies: orthoretroviridae and spumaretroviridae. The subfamily orthoretroviridae further consists of six genera (alpharetrovirus, betaretrovirus, gammaretrovirus, deltaretrovirus, epsilonretrovirus, lentivirus) whereas the spumaretroviridae only has one genus (spumavirus) [57]. Each genus is further subdivided into different types of subspecies. Lentiviruses belong to the subfamily orthoretroviridae and are single‐stranded, positive‐sense RNA viruses that have a diploid genome with a length of 7‐12kb. The virions are enveloped and 80‐100nm in diameter. Moreover, these viruses are characterized by two viral enzymes: reverse transcriptase and integrase, which convert the RNA genome into double‐stranded DNA and integrate it into the host cell genome. Another characteristic of lentiviruses is their ability to infect both dividing and non‐dividing cells. Infection with lentiviruses can result in different diseases such as immunodeficiencies, anemia and encephalitis [58]. Since in this dissertation a HIV‐1‐ derived lentiviral vector was used, further descriptions are based on the HIV‐1 genome.
1.4.2 The lentiviral genome
The lentiviral genome exists as two single‐stranded positive‐strand RNA molecules and has a 5’ cap structure and a 3’ polyadenylation. The proviral genome is flanked by two, 640bp long terminal repeats (LTR) that are required for transcription, reverse transcription and integration into the host cell genome. Each LTR is made of a U3, R and U5 region and contains several promoter and enhancer elements. The lentiviral genome further contains a primer binding site (PBS) necessary for binding of a tRNA that initiates the reverse transcription of the RNA genome. Located in 3’ to the PBS is the packaging and dimerization signal psi (Ψ), which mediates packaging of the RNA genome during virion assembly. At the 3’ LTR, the polypurin tract (3’PPT) required for the synthesis of the second DNA strand during reverse transcription is encoded. The viral core proteins, enzymatic proteins, and the envelope glycoproteins shared by all retroviruses are encoded in the gag, pol and env gene. Complex retroviruses, such as lentiviruses, also contain six additional genes (tat, rev, vif, vpr, vpu, and nef) important for HIV infection and pathogenicity, making it a total of nine open reading frames (Figure 6) [58]. 13 INTRODUCTION
A HIV RNA Genome
PBS cPPT RRE 3’PPT
RU5 Ψ U3 R Reverse Transcription
B Viral DNA Genome nef tat
gag vif rev LTR LTR
U3 RU5 pol vpr env U3 RU5 vpu
Figure 6: The RNA and proviral genome of HIV‐1. (A) Structure of the RNA genome. The 5’ and 3’ end are modified, carrying a cap structure and a polyadenylation, respectively (not shown). At the primer binding site (PBS) a tRNA for the initiation of reverse transcription is bound (not shown). Psi (Ψ) is the packaging signal, U3, R and U5 mark the end of the HIV genome and form the long terminal repeats (LTR) after reverse transcription. The central polypurine tract (cPPT) is important for transportation of the provirus into the nucleus. (B) Schematic of the proviral genome. During reverse transcription, the U3 region at the 3’ end is copied to the 5’ end and the 5’ U5 region to the 3’ end to form the long terminal repeats. The location of the nine lentiviral genes is shown.
The gag gene encodes four viral capsid proteins that are derived from a single precursor Gag protein called p55 and that are necessary for the formation and release of viral particles. For lentiviral replication, the viral protease, reverse transcriptase and the integrase are required, which are encoded in the pol gene. The gag and pol genes are synthesized from the same mRNA transcript by a ribosome frameshifting near the 3’ end of the gag gene. The resulting Gag‐Pol precursor protein is processed by the viral protease inside the virions. The env gene encodes the viral surface protein gp120 and transmembrane gp41 protein. Similar to the Gag and Pol proteins, these mature proteins are processed from a single precursor protein, gp160, by proteolytic activity. The Gag, Pol and Env proteins are essential for the formation of mature virions and the infection of target cells. One of the accessory proteins encoded in the genome of lentiviruses is the transactivator Tat, which binds to the Tat response element (TAR) located in the LTR and plays a critical role for the transactivation and transcription of proviral DNA as well as the lentiviral replication. Besides, the Rev protein promotes the nuclear export of lentiviral RNA by binding specifically to the Rev response element (RRE) located in the env region, thereby increasing the half‐life of viral mRNAs. Both Tat and Rev are translated from early transcribed mRNA [58].
1.4.3 The life cycle of lentiviruses
The life cycle of lentiviruses is divided into early and late stages. In the early stages, virus entry, reverse transcription of the RNA genome and integration into the genome occur. The late stages of the lentiviral life cycle are characterized by the transcription of the integrated viral genome, translation of viral transcripts into proteins required for viral assembly and budding from the host cell (Figure 7). 14 INTRODUCTION
The lentiviral life cycle begins with the recognition and binding of the viral surface protein Gp120 to its primary receptor CD4 and the co‐receptor CXCR4 on T‐lymphocytes or CCR5 on macrophages and dendritic cells. Binding initiates a conformational change in the viral transmembrane protein Gp41 required for the fusion of the viral membrane with the cell membrane. Upon viral entry into the cell, the capsid proteins are uncoated, and the lentiviral genome, reverse transcriptase, protease and integrase are released into the cytoplasm. Subsequently, the RNA genome is converted into the complementary double‐stranded DNA by the enzymatic activity of the reverse transcriptase [58]. The viral enzyme integrase (IN) specifically binds to the U3 region in the 5’ LTR and the U5 region in the 3’LTR of the proviral cDNA. IN removes nucleotides from the 3’ ends of the viral DNA beyond a conserved ‘CA’ dinucleotide, thereby creating two single‐stranded 5’ overhangs. The processed proviral DNA is subsequently imported into the nucleus, where the cellular DNA is cleaved by IN, the 5’ dinucleotide overhang in the viral DNA is removed and the 3’ DNA ends joined to the 5’ ends of the genome [59]. Following stable integration, the host cell transcriptional machinery is hijacked to transcribe the provirus. The 5’ LTR has basal promoter activity that is sufficient to initiate transcription by cellular RNA polymerase II. The first mRNA transcripts are multi‐spliced mRNAs that are exported from the nucleus into the cytoplasm where they are translated into Rev, Tat and Nef proteins, required for the initiation of subsequent events in viral transcription. The Tat protein binds to the TAR element at the 5’ end of viral mRNAs, leading to the transactivation and transcription of other genes from the viral DNA. In addition to the Tat protein, the Rev protein binds to the Rev‐responsive element on viral mRNA transcripts and removes them from the splicosome, which leads to the export of singly‐spliced and unspliced mRNA and viral genomes into the cytoplasm. Singly‐ spliced mRNAs are translated at the ribosomes into the proteins Env, Vif, Vpr, and Vpn, whereas unspliced mRNA are translated into the Gag‐Pol precursor protein. Following transcription and translation, the viral proteins and the genomic RNA assemble at the cell membrane, where viral Env proteins are integrated into the cell membrane. The unspliced RNA genome copies are then packaged into the virions and released from the cell. Subsequently, the multimerization of Gag and Gag‐Pol precursor proteins activates the viral protease, converting immature to mature HIV virions (Figure 7) [58]. 15 INTRODUCTION
Figure 7: Lentiviral life cycle. In the early stage of viral infection, the virus attaches to the target cell and fuses with the cellular membrane to release the capsid into the cell. The capsid is subsequently uncoated, the RNA genome reverse transcribed, imported into the nucleus and integrated into the genome. After integration, the lentiviral genes are transcribed and exported into the cytoplasm. The viral mRNA is translated, and the lentiviral RNA genome and the proteins assemble into virions at the cell membrane. NPC: nuclear pore complex. Image taken from [60]
The integration profile of HIV in the genome of their host cells is non‐random. HIV preferentially integrates into genomic regions with a high gene density and particularly into coding sequences. However, integration upstream of the transcription start site is not favored [61, 62]. Furthermore, HIV integration is influenced by the transcriptional activity: the integration frequency into transcriptionally active genes is increased [63], but genes with high transcriptional activity are less favored [61]. Regions rich in the dinucleotide CpG, termed CpG islands, commonly correspond to gene regulatory regions that frequently contain promoter and enhancers. For HIV, these regions and their surroundings are disfavored [61]. Hence, gene‐dense and transcriptionally‐active genomic regions, but not CpG islands are preferred sites of HIV integration.
1.4.4 Structure of lentiviral vectors for gene therapy
Because of their ability to infect both dividing and non‐dividing cells as well as stably integrate into the genome, lentiviruses are used as vehicles to introduce transgenes into target cells for more than 20 years [58]. In order to generate replication‐incompetent viruses for gene transfer, the lentiviral genome sequence was modified. Today, plasmid vector systems are used in which all elements required for production of functional virions (trans‐acting factors) of the viral genome are divided onto four plasmids to reduce the likelihood of recombination and formation of replication‐competent viruses. In principal, the first plasmid encodes the packaging genes gag and pol. In the second plasmid, the rev gene that is required during lentiviral transcription 16 INTRODUCTION
is encoded. On the third plasmid, the Env protein is often replaced by the VSV‐G protein (vesicular stomatitis virus ‐ protein G) to increase the tropism of the lentivirus. In order to prevent that the three plasmids containing the trans‐acting factors are packaged into mature virions and that replication‐competent viruses form, the packaging signal psi and the LTR sequences have been cloned into a fourth plasmid, called the transfer vector. This vector is packaged into the virions, and contains the sequence for reverse transcription and the transgene. Moreover, to further reduce the likelihood of recombinations, the U3 region in the LTR containing the promoter and enhancer elements is truncated [64]. Furthermore, the requirement of lentiviral replication for the Tat protein was eliminated, by replacing the U3 region in the 5’ LTR with a constitutively active promoter from the cytomegalovirus (CMV) or RSV. Vectors with these LTR modifications are called self‐inactivating (SIN), because they lack LTR promoter activity following integration, thereby reducing the likelihood of transcription activation of proto‐oncogenes located in vicinity to the lentiviral integration site. Moreover, increased transduction efficiency and transgene expression in vitro and in vivo was achieved by placing a central polypurine tract (cPPT) sequence downstream of the RRE in the transfer vector. Another genetic element in the transfer vector is the posttranscriptional regulatory element WPRE (woodchuck hepatitis virus post‐transcriptional element), which increases the amount of transgenic mRNA. For production of functional viral particles, the four plasmids are co‐ transfected into a packaging cell line in which the genomic information is transcribed and translated into mature viral particles, containing the transfer vector.
In addition to integration‐competent lentiviruses, episomal‐remaining, non‐integrating lentiviral vectors have been developed, which enable transient transgene expression in infected cells. These lentiviruses are termed integrase‐deficient lentiviruses (IDLV) and are generated by introducing a point mutation at the catalytic core of the integrase gene that changes the amino acid at position 64 from aspartic acid to valine (D64V). This modification was shown to inhibit integration by up to four logs compared to integrase‐competent lentiviruses [65, 66], and several studies suggest that the few observed integration events are mediated by cellular DNA repair mechanisms [65]. Indeed, integrated IDLVs showed a high frequency of LTR deletions, a key signature of NHEJ‐repair activity, and occurred at sites of DNA damage and repair. Additionally, blocking ATM, a key player in DNA repair, also prevented integration of IDLV at DSB sites [67].
1.5 Scientific aims
Current methods for DSB detection have several limitations, which hamper the analysis of induced and repaired DSB as well as genomic instabilities in irradiated cells: 1) Immunostaining of DNA repair proteins is an indirect detection method that does not enable the identification of DSB sites at single‐nucleotide resolution, 2) DNA repair proteins form microscopically‐visible foci which disassemble upon DNA repair and do not allow DSB site analysis in surviving cell populations, 3) Immunostaining does not give any information on the relationship between DSB locations, DNA repair pathway choice and cell fate decision. Even though the high‐throughput sequencing platforms are frequently used to study the mutational spectra in various cancer types [68], the high costs for sequencing large numbers of whole genomes and the analysis of sequencing data from heterogeneous cell populations make this method impractical for the identification of low frequency genomic events. Moreover, it does not provide any information, whether the identified mutation has functional consequences on cell survival and therapy resistance. Hence, there is little information on how radiation‐induced and repaired DSB sites are distributed, how radiation‐induced DNA damage is being survived and how these damages induce radioresistance and cell transformation. Analyzing the genomic distribution of radiation‐induced DSB in irradiated and expanded cells at single‐nucleotide resolution will help to improve our understanding of the 17 INTRODUCTION
mechanisms of radiation‐induced genomic instability and radiation resistance. In this thesis, a new methodology to capture induced and repaired DSB sites in radiation‐surviving cell populations has been used to study DSB site distribution. Target cells become transduced with an integrase‐deficient lentivirus (IDLV) that carries a point mutation in its integrase gene, preventing integrase‐mediated integration of the proviral DNA into the host genome of the cell. Upon DSB induction by irradiation, these IDLV DNA molecules serve as molecular tags, which become stably integrated into the genome by the cellular NHEJ‐repair activity, thereby marking the DSB site in vivo. The integration of IDLV at induced and repaired DSB sites can be followed by PGK promoter‐driven EGFP expression from integrated IDLV by flow cytometry. The DSB repair sites can be amplified and identified using LAM‐PCR and deep sequencing (Figure 8).
IR/Dox IR/Dox IR/Dox
Genomic DNA DSB DSB DSB Induced DSB IDLV NHEJ Repair of DSB
NHEJ
LTR IDLV-PGK-eGFP LTR IDLV “DSB trapping”
5’LAM-PCR 3’LAM-PCR Amplification of DSB by LAM-PCR
GTACCTGTTCA TCTGGAAGCTATTT Localization of DSB at Sequencing of DSB Loci single nucleotide resolution
Figure 8: Proposed mechanism of IDLV‐based DSB capturing. The double‐stranded DNA bait delivered into cells by IDLV is captured by the cellular NHEJ‐repair machinery at DSB sites. This leaves a stable genetic tag which enables the mapping and tracking of radiation (IR)‐ or doxorubicin (Dox)‐induced and repaired DSB in the genome of treated cells. The frequency of DSB tagging can be followed by PGK promoter‐driven EGFP expression during expansion. Localization of captured DSB is performed by amplifying the vector‐genome junction using 5’ and 3’ LAM‐PCR and deep sequencing. DSB: DNA double‐strand breaks, IR: irradiation; Dox: doxorubicin; IDLV: integrase‐deficient lentivirus; NHEJ: Non‐Homologous End Joining, eGFP: enhanced green fluorescent protein; PGK: Phosphoglycerate Kinase 1 promoter; LTR: Long terminal repeat
The following scientific aims were addressed:
1. Can radiation‐induced DSB be stably marked, tracked and identified at single‐nucleotide resolution during expansion? 2. How do transcriptional and epigenetic states influence DSB site distribution? 3. Do frequently damaged and repaired genomic regions exist that are likely to influence radiation‐ induced genomic instability and radiotherapy?
The obtained results from this work should bring new insights into the mechanisms that initiate DSB‐induced genomic instability, radiotherapy resistance and carcinogenesis.
18 MATERIALS AND METHODS
2. MATERIALS AND METHODS
2.1 Materials
2.1.1 Chemicals
Chemicals/Reagents Company 100 bp / 1 kb Marker Invitrogen Agarose LE Sigma Ampicillin Roth
Aqua ad iniectabilia (dH2O) Braun BD™ Cytometer Setup & Tracking Beads Kit BD Beckton Dickinson BD FACS Clean Solution BD Beckton Dickinson BD FACS Flow Sheath Fluid BD Beckton Dickinson BD FACS Rinse Solution BD Beckton Dickinson Bovine Serum Albumine (BSA) Sigma Bromphenol blue Sigma Deoxyribonucleotic triphosphate (dNTP) Fermentas (Thermo Fisher Scientific) Dimethylformamid Sigma Dulbecco's Modified Eagle Medium (DMEM) Invitrogen Ethanol VWR Ethidiumbromid solution (0.07%) Applichem Ethylenediaminetetraacetic acid (EDTA) Applichem Fetal Calf Serum (FCS) PAN Glycerol Sigma‐Aldrich Guanidin hydrochloride Sigma Hexanucleotid Mix (10x) Roche Human genomic DNA Roche Iscove's Modified Dulbecco's Medium (IMDM) Invitrogen Isopropanol Sigma‐Aldrich LB Agar Miller US Biological Lithiumchlorid (LiCl) Sigma Loading Buffer (5x) Elchrom Scientific Luria‐Bertani Broth (LB) Invitrogen
Magnesiumchlorid (MgCl2) Sigma ms2RNA Roche Sodium hydroxid (NaOH) Fluka Dynabeads M‐280 Streptavidin Dynal Dynabeads MyOne Streptavidin T1 Dynal PCR Grade Water Roche Penicillin/Streptomycin (Pen/Strep) Invitrogen Dulbecco’s Phosphate Buffered Saline (DPBS; pH 7,4) Gibco Polybrene (1000x; 8 µg/ml) Sigma Polyethylenimin (PEI; 1mg/ml) Sigma‐Aldrich Propidium iodide (1mg/ml) Molecular Probes (Invitrogen) Proteinase K Roche/Qiagen Puromycin Invitrogen
RNase/DNase free H2O Ambion Tris‐borate‐EDTA (TBE) Buffer (10x) Amresco Trizma‐HCl (Tris‐HCl) Applichem 19 MATERIALS AND METHODS
Trypanblue Invitrogen Trypsin/EDTA (0,05%) Life Technologies Tween 20 Sigma Vectashield Mounting Medium with DAPI Vector Laboratories X‐Gal Sigma
2.1.2 Enzymes
Enzymes Company CircLigase ssDNA ligase Epicentre Klenow DNA Polymerase Roche Quick DNA ligase Epicentre T4 DNA Ligase New England Biolabs Restriction enzymes with respective buffers New England Biolabs SYBRGREEN I Mix Roche Taq DNA Polymerase Qiagen/Genaxxon
2.1.3 Bacteria
Bacterial Strain Genotype Company E.coli TOP10 F‐ mcrA Δ(mrr‐hsdRMS‐mcrBC) Φ80lacZ ΔM15 Life Technologies ΔlacX74 recA1 araD139 Δ(ara leu) 7697 galU galK rpsL (StrR) endA1 nupG ‐ ‐ E.coli Stbl3 F mcrB, mrr‐hsdS20 (rB , mB‐), recA13, supE44 Life Technologies ara‐14, galK2, lacY1, proA2, rpsL20(StrR), xyl‐5 λ‐leumtl‐1
2.1.4 Cell lines and primary cells
Name Description Company A549 Human alveolar adenocarcinoma cell line [69] ATCC HEK 293T Human embryonic kidney cell line, stably ATCC expressing the SV40 T‐antigen [70] Hela Human epithelial carcinoma cell line ATCC NHDF‐A Adult normal human dermal fibroblasts PromoCell PC3 Human prostate cancer cell line [71] ATCC U87 Human glioblastoma cell line [72] ATCC
20 MATERIALS AND METHODS
2.1.5 Antibodies
Name Description Company Alexa Fluor® 647 anti‐ anti H2AX‐specific antibody labeled Biolegend H2A.X‐Phosphorylated with Alexa Fluor 647 (Ser139) Antibody (250ng/ml) Alexa Fluor® 647 Mouse Isotype control antibody, Antibody Biolegend IgG1, κ Isotype Ctrl (ICFC) labeled with Alexa Fluor 647
2.1.6 Plasmids
Name Description Lentiviral transfer vector expressing EGFP under control of pCCLsincPPT.PGK‐IRES‐eGFP.WPRE (LV106) human PGK promoter Plasmid encoding gag/pol genes and carries D64V mutation in LV001 integrase gene LV102 env‐encoding plasmid LV103 Plasmid encoding the VSV‐G protein #1211 Plasmid encoding CCR5‐specific ZFN monomer 1 #1212 Plasmid encoding CCR5‐specific ZFN monomer 2 pUC‐derived subcloning vector with multiple cloning site, pCR2.1 TOPO‐TA ampicillin and kanamycin resistance cassette
2.1.7 Oligonucleotides
DNA oligonucleotides were partially biotinylated (B) at the 5‘ terminus. Additional modifications include dideoxycytosine (ddC) to inhibit ligation of DNA oligos to 3’ terminus and phosphothiorate (*) in the phosphate backbone of the last three bases of some DNA oligos. Oligonucleotides were cleaned up by high performance liquid chromatography (HPLC) and lyophilized. Oligonucleotides were purchased from MWG or Sigma Aldrich. B,
Biotin; D: degenerated base; LC, Linker cassette; Tit, Titanium 454 primer; (N)2‐6, Recognition sequence for pyrosequencing.
2.1.7.1 Standard primers for q‐RT‐PCR
Name Sequence (5´‐3´) Myo461‐439 CTCCCAGTGGCACAGCAGTTAGG Myo122‐143 TGTGCCCCAGGTTTCTCATTTG GFP2 fwd TGAGCAAGGGCGAGGAGCTGTT GFP3 rev GCCGGTGGTGCAGATGAACT
21 MATERIALS AND METHODS
2.1.7.2 Primers used for linker cassettes in LAM‐PCR
Name Sequence (5´‐3´) LC1 GACCCGGGAGATCTGAATTCAGTGGCACAGCAGTTAGG LC1 (CATG) GACCCGGGAGATCTGAATTCAGTGGCACAGCAGTTAGGCATG LC3 (CG) CGCCTAACTGCTGTGCCACTGAATTCAGATC LC2 CCTAACTGCTGTGCCACTGAATTCAGATC LC3 (AATT) AATTCCTAACTGCTGTGCCACTGAATTCAGATC LC3 (TA) TACCTAACTGCTGTGCCACTGAATTCAGATC
2.1.7.3 Primers used for LAM‐PCR
Name Sequence (5´‐3´) SK‐LTR 1bio B ‐GAGCTCTCTGGCTAACTAGG SK‐LTR 3bio B ‐GAACCCACTGCTTAAGCCTCA SK‐LTR 4bio B ‐AGCTTGCCTTGAGTGCTTCA SK‐LTR 5 AGTAGTGTGTGCCCGTCTGT SK‐LTR 5 1/2 GTGTGACTCTGGTAACTAGAG LCI GACCCGGGAGATCTGAATTC LCII GATCTGAATTCAGTGGCACAG
2.1.7.4 Fusionprimer for Pyrosequencing
Name Sequence (5´‐3´)
Tit3nrLV CCATCTCATCCCTGCGTGTCTCCGACTCAG(N)6‐10 GATCCCTCAGACCCTTTTAGTC Tit3SKLV CCATCTCATCCCTGCGTGTCTCCGACTCAG(N)6‐10 TGTGTGACTCTGGTAACTAG
Tit5SKLV CCATCTCATCCCTGCGTGTCTCCGACTCAG(N)6‐10 AAGCAGATCTTGTCTTCG
MegaL6/10T GCCTCCCTCGCGCCATCAG(N)6‐10GATCCCTCAGACCCTTTTAGTC
MegaL_U3_10/T GCCTCCCTCGCGCCATCAG(N)6‐10AAGCAGATCTTGTCTTCG MiS3nrLV AATGATACGGCGACCACCGAGATCTACACTCTTTCCCTACACGACGCTCTTCCGATCT
(N)6‐10GATCCCTCAGACCCTTTTAGTC AATGATACGGCGACCACCGAGATCTACACTCTTTCCCTACACGACGCTCTTCCGATCT Mis3SKLV (N)6‐10TGTGTGACTCTGGTAACTAG Tit Linker B‐CCTATCCCCTGTGTGCCTTGGCAGTCTCAGAGTGGCACAGCAGTTAGG Megalinker B‐GCCTTGCCAGCCCGCTCAGAGTGGCACAGCAGTTAGG
MiS_LK GACCCGGGAGATCTGAATTCAGTGGCACAGCAGTTAGG(N)16CTA
22 MATERIALS AND METHODS
2.1.7.5 Primers and oligos used for direct DSB labeling approaches
Name Sequence (5´‐3´) L.CDSB 1 B‐CCTTGGGAGGGTCTCCTCTGAGTGATTGACAAAAAD L.CDSB 2 B‐CTTGGGAGGGTCTCCTCTGAGTGATTGACAAAAA L.CDSB 2‐1 CTTGGGAGGGTCTCCTCTGAGTGATTGACAAAAA E.CDSB 3 CTTGGGAGGGTCTCCTCTG P‐GTAGTCAATACATCAGAGGAGACCAGTCGGATGCAACTGCAAGATATTG dsGRUV1 GATACACGGGTACC*C*G*G P‐GTAGTCAATACATCAGAGGAGACCAGTCGGATGCAACTGCAAGATATTG dsGRUV2 GATACACGGGTACCCGGGACCCG*G*G*A P‐CCGGGTACCCGTGTATCCAATATCTTGCAGTTGCATCCGACTGGTCTCCT dsGRUVrev CTGATGTATTGA*C*T*A‐ddC P‐GTAGTCAATACATCAGAGGAGACCAGTCGGATGCAACTGCAAGATATTG ssGRUV5‐A‐Phthio GATACACGGGTACC*C*G*G‐ddC GRUV5‐2expo CATCCGACTGGTCTCCTCTGATG
2.1.8 Commercial kits
Kits Company Ampure XP beads Beckman Coulter Fast‐Link Ligation Kit Epicentre Fluorescence Biotin Quantitation Kit ThermoScientific High Pure PCR Template Preparation Kit Roche HiSpeed Plasmid Purification Kit Qiagen Microcon‐30 Kit Millipore QIAamp DNA Mini Kit Qiagen QIAfilter Plasmid Purification Kit Qiagen QIAprep Spin Miniprep Kit Qiagen QIAprep Spin Maxiprep Kit Qiagen QIAquick PCR Purification Kit Qiagen QIAquick Gel Extraction Kit Qiagen TOPO‐TA Cloning Kit Invitrogen
23 MATERIALS AND METHODS
2.1.9 Buffers, Media, Solutions
Buffers, Media, Solutions Chemical Final Concentration 3M LiCl Solution** Tris‐HCl (pH 7.5) 10 mM EDTA 1 mM LiCl 3 M 6M LiCl Solution** Tris‐HCl (pH 7.5) 10 mM EDTA 1 mM LiCl 6 M Acid‐isopropanol solution HCl (1N) 0.04N Isopropanol Ampicillin** Ampicillin sodium salt 100mg/ml Desalted water Cell culture freezing medium DMSO 15% FCS 30% Cell culture medium 55% Denaturation solution NaOH 0.1 M DMEM Cell culture medium DMEM 89% FCS 10% Pen/Strep 1% DNA loading buffer (5x) Tris‐HCl (pH 7) 25 mM EDTA (pH 8) 150 mM Bromphenol blue 0.05% Glycerol 25% DPBS‐Tween 20 (DPBST) Tween 20 0.2% DPBS (pH 7.4) IMDM cell culture medium IMDM FCS 10% Pen/Strep 1% LB‐Agar* LB‐Miller Agar 3.7% Desalted water
LB‐medium* LB 2% Desalted water MTT stock solution** MTT 5mg/ml 1xDPBS Paraformaldehyde solution** Paraformaldehyde 4% DPBS (pH 7.4) Triton solution Triton X‐100 0.5% DPBS (pH 7.4) Washing solution for paramagnetic beads** DPBS (pH 7.4) BSA 0.10% X‐Gal X‐Gal 20 mg/ml Dimethylformamid * Autoclaved and addition of antibiotics (Ampicillin, Kanamycin) after cooling to 60°C.
** Solutions were sterilized by filtering through 0.22µm filter
24 MATERIALS AND METHODS
2.1.10 Disposables
Disposable Company Sheeting VWR Sheeting for LightCycler LC480 Roche Polypropylene Round Bottom Tube (15 ml) BD Beckton Dickinson Polystyrene Round Bottom Tube (5 ml) BD Beckton Dickinson Lid for 96‐well plates Greiner bio‐one Falcon, BD™ Falcon™ Tubes (15 ml, 50 ml) BD Beckton Dickinson Filter (0,22 µm, 500 ml) Millipore Filter System (0,22 μm) Stericup® Millipore Gloves (Nitril) Microflex Ionoculation Loop Sarstedt Parafilm Brand PCR‐Tubes Softtubes (0,2 ml) Kisker Biotech Pipette tips Starlab Pipette tips for LightCycler LC480 Starlab Mixing plates (96‐well) Greiner bio‐one Plates for LAM‐PCR (96‐well) Greiner bio‐one Plates for LightCycler LC480 (96‐well) Roche Photobase paper VM 65 H Mitsubishi Reaction tubes (0,5 ml, 1,5 ml, 2 ml) Eppendorf RNase/DNase‐free reaction tubes Ambion (0,5 ml, 1,5 ml, 2 ml) Ultracentrifugation tube Beckman Coulter Cell culture flasks (25 cm2, 75 cm2, 225 cm2) Nunc Brand Products Cell culture pipettes (2‐50 ml) BD Beckton Dickinson Cell culture plates (6‐, 12‐, 24‐, 96‐well) BD Beckton Dickinson Cell culture dish (10 cm, 15 cm) Greiner bio‐one Cell scraper BD Beckton Dickinson Cryotube BD Beckton Dickinson
2.1.11 Equipment
Equipment Company Autoclave Systec BDTM LSRII Flow cytometer BD Beckton Dickinson Cell culture hood Heraeus Centrifuges Eppendorf Lysine‐coated coverslips BD Beckton Dickinson Electrophoresis Power Supply Pharmacia/Elchrom Scientific Fluorescence microscope Carl Zeiss Jena Freezer ‐20°C Liebherr Freezer ‐80°C Sanyo Fridge Liebherr Gel documentation system Peqlab Gel electrophoresis chamber Biometra 25 MATERIALS AND METHODS
Heating Block Eppendorf Horizontal shaker (KS 250B) IKA Labortechnik Incubator (37°C) Binder Lifesciences Ice machine Ziegra Infinite 200” plate reader Tecan LightCycler LC480 Roche Liquid Nitrogen Tank Thaylor‐Wharton Magnetic Particle Collector MPC 96 Dynal Microplate reader Biotek Microscope Carl Zeiss Jena Microwave Siemens Multi‐channel pipette Eppendorf NanoDrop spectrophotometer Thermo Scientific Neubauer counting chamber Optik Labor OptimaTM L‐90K Ultrazentrifuge Beckman Coulter PCR cycler Biometra Picofuge NeoLab Pipettes (Pipetman P2, P10, P200, P1000) Eppendorf Pipette device Integra Bioscience Pipetboy acu Proton and Carbon Radiation Source Heidelberg Ion Therapy Center Scales Sartorius Special accuracy weighing machine Sartorius Vacuum pump NeoLab Videoprinter Mitsubishi Vortexer (MS1) IKA Labortechnik Water bath Thermo Electron Coporation XRAD320 X‐ray device Precision X‐Ray
2.1.12 Software and data bases
Software and data bases Company Adobe Acrobat X Pro Adobe BLAST Search http://www.ncbi.nlm.nih.gov/blast BLAT Search http://genome.uscs.edu/ FACSDiva Software V6.1.2 BD Beckton Dickinson Fiji image processing package http://www.fji.sc/ Galaxy project in house USCS Genome Browser http://genome.uscs.edu/ High‐throughput Insertion Site Analysis Pipeline (HISAP) in house Ingenuity Pathway Analysis (IPA) http://www.ingenuity.com/ Lasergene DNA Star LightCycler LC480 Software Roche Office 2007 (Word, Excel, PowerPoint) Microsoft Photoshop CS5.1 Adobe R‐Program (2.13.1) cran.r‐project.org
26 MATERIALS AND METHODS
2.2 Methods
2.2.1 Cell Culture Methods
2.2.1.1 Cell Cultivation
A549 (human alveolar basal epithelial cancer cell line), PC3 (human prostate cancer cell line) and U87 (human glioblastoma cell line) were maintained with twice‐weekly subculture in Dulbecco`s Minimal Essential Medium (DMEM) supplemented with 10% fetal calf serum (FCS) and 1% Penicillin/Streptomycin (Pen/Strep). 293THEK (human embryonic kidney cell line) were maintained in IMDM supplemented with 10% FCS and 1% Pen/Strep. Adult normal human dermal fibroblasts (NHDF‐A) were cultivated in DMEM supplemented with 10% FCS, 1% Pen/Strep and 1% L‐Glutamine. NHDF‐A cultivated for more than nine passages were not used for radiation and
chemotherapy experiments. All cells were cultured at 37°C and 5% CO2. For passaging, the adherent cells were washed once with pre‐warmed 1xDPBS and subsequently detached using 3ml or 5ml 0.05% trypsin‐EDTA for a 10cm or 15cm cell culture dish, respectively. The cells were incubated at 37°C in an incubator for 7 minutes. Trypsin activity was blocked by addition of 7ml (10cm cell culture dish) or 10ml (15cm cell culture dish) cell culture medium supplemented with 10% FCS.
2.2.1.2 Freezing and thawing of cells
In order to freeze cells, cell culture freezing medium containing 3ml DMSO, 6ml FCS and 11ml cell culture medium, was prepared. Cells were detached from the cell culture dish, centrifuged at 1,000rpm and room temperature for 5min, and resuspended in cell culture medium. Subsequently, 400ul aliquots were transferred into cryotubes. To each tube, 400µl freezing medium was added, and the tubes placed in freezing boxes. The freezing boxes were stored at ‐80°C over night. Finally, cryotubes were placed in a liquid nitrogen tank. For thawing, cryotubes containing cells were incubated in a water bath for two minutes until they were thawn. Then, cells were resuspended in 10ml cell culture medium in 15ml reaction tubes, centrifuged at 1000rpm for 5min and seeded in cell culture dishes of appropriate size.
2.2.1.3 Cell Counting
For cell counting, 10µl medium containing cells detached from the cell culture dish were transferred to a 1.5ml reaction tube and diluted 1:10 in 90µl trypan blue. From this mixture, 10µl were loaded into a Neubauer counting chamber and the average number of cells in the chamber calculated accordingly:
. / 10 10
27 MATERIALS AND METHODS
2.2.1.4 Transfection
The term transfection describes several processes for the introduction of nucleic acids into mammalian cells. In this work, transfection was carried out by mixing the cationic lipid polymer poly‐ethylene inositol (PEI) with the DNA to produce liposomes, which fuse with the cell membrane to release the DNA into the cell.
Transfection of lentiviral plasmids for virus production was performed on 293T human embryonic kidney (HEK) cells (chapter 2.2.1.6). Therefore, 1x106 293T HEK cells were seeded in a 15cm cell culture dish with 15ml IMDM 24h prior to transfection. In total, 20 dishes were used for virus production. Following incubation, the cell culture medium was removed and replaced with fresh 10ml IMDM supplemented with 15% FCS.
Reagent Volume for one 15cm dish PEI [1mg/ml] 167.25µl PEI solution IMDM without FCS ad 2.5ml LV001 [1mg/ml] 12.5µg LV102 [1mg/ml] 6.25µg DNA solution LV103 [1mg/ml] 9µg LV106 [1mg/ml] 28µg IMDM without FCS ad 2.5ml
For transfection of CCR5‐specific ZFN (kindly provided by Dr. Richard Gabriel), 1x106 A549 or 293T HEK cells were seeded in a 15cm cell culture dish with 15ml IMDM 24h prior to transfection. Following incubation, the cell culture medium was removed and replaced with fresh 10ml IMDM supplemented with 15% FCS.
Reagent Volume for 15cm dish PEI [1mg/ml] 18µl PEI solution IMDM without FCS ad 2.5ml #1211 [1mg/ml] 3µl DNA solution #1212 [1mg/ml] 3µl IMDM without FCS ad 2.5ml
The Zinkfinger nuclease consists of two monomers (#1211, #1212) which dimerize inside the nucleus at the target DNA sequence to induce a DSB. Each monomer has a DNA binding domain which can be targeted basically against every locus in the genome and a FokI‐nuclease domain that introduces the DSB upon dimerization.
The transfection solutions were mixed together and incubated at room temperature for 30 minutes prior to addition to the cell culture dish. Then, the cells were incubated with the transfection mixture over night. The next morning, the medium was discarded and replaced with fresh 12.5ml IMDM (10% FCS).
2.2.1.5 Transduction
The introduction of DNA into target cells by viral delivery systems is termed transduction. Here, lentiviruses carrying the IDLV vector were used to transduce target cells. During transduction, the lentiviruses bind to and fuse with the cell membrane. The viral capsid enters the cytoplasm, where the nucleic acid is released. The lentiviral RNA is reverse transcribed into double‐stranded DNA and transported into the nucleus. 28 MATERIALS AND METHODS
For transduction, 1x106 cells in 15ml medium were plated in 15cm cell culture dishes 24h prior to transduction. Then, the cell culture medium was removed and replaced with fresh 10ml medium containing 1x polybrene. Polybrene increases the transduction efficiency by neutralizing the electrostatic charges between the virion and the negative cell surface. The virus was pooled in a single reaction tube, added to the cells and replaced with fresh medium after 24h incubation.
2.2.1.6 Virus production
For virus production, 293T HEK cells were co‐transfected (chapter 2.2.1.4) with the four plasmids encoding the Gag, Pol (LV001), Rev (LV102) and VSV‐G (LV103) proteins as well as the transfer vector carrying the eGFP gene (LV106). Following 48h incubation, the virions were isolated from the cell culture medium. The medium was removed from the cell culture dish and filtered through a 0.22µm filter. The flowthrough containing the virions was then split into six vials and concentrated by ultracentifugation at 20°C and 20,000 x g for 2h. The pellet in each vial containing the virions was resuspended in 70µl 1xDPBS, incubated for 20 minutes at RT, and the virions pooled in a single 1.5ml reaction tube. The tube was placed in a rotator for 20 minutes and finally stored in 10µl aliquots at ‐80°C.
2.2.1.7 Determining the lentiviral titer on Hela cells
In order to determine the efficiency of the virus production and to calculate the multiplicity of infection (MOI) for transduction of the target cells, the titer of the lentivirus production was determined. The MOI is the number of infectious particles that can infect a specific number of cells: MOI1 means that for a defined number of cells the same number of virus particles is present in the solution.
Since the IDLV vector used in this study encodes EGFP under control of the human PGK promoter, the lentiviral titer was determined by FACS analysis. First, Hela cells were seeded in all wells of a 6‐Well plate at a density of 5 x 104 cells per well and incubated for 24h at 37°C, allowing each cell to divide once. The medium was then removed and replaced with 500µl DMEM containing 2µl 1000x polybrene. A virus dilution series in 1ml DMEM was prepared and 500µl transferred to each well containing cells. The cells were incubated with IDLV for 48h before FACS analysis. To calculate the number of infectious particles (Titration units, TU) per ml the following formula was used:
% 100 10
2.2.1.8 MTT Assay
The MTT assay is a quantitative colorimetric assay for mammalian cell survival and proliferation [73]. The tetrazolium salt MTT (3‐(4,5‐dimethylthiazol‐2‐yl)‐2,5‐diphenyl tetrazolium bromide) is a yellow substance that is metabolized into purple, insoluble salt crystals by dehydrogenase enzymes in mitochondria of living cells exclusively. Thus, the degree of crystal formation is proportional to the number of living cells. 29 MATERIALS AND METHODS
The MTT reaction solution (10µl MTT in 100µl medium) was added to all assay wells containing cells in a 96‐well plate, and plates were placed in the incubator at 37°C for 4h. Subsequently, the MTT‐reaction solution was removed, replaced with 100µl acid‐isopropanol solution (0.04N HCl in isopropanol) and mixed thoroughly to dissolve all crystals. The solution was incubated at RT for 5min, and finally the absorption measured in a microplate reader at wavelength of 610nm. Triplicate measurements for each substrate concentration were performed. Here, the MTT assay was used to determine the LD50 values of etoposide, doxorubicin, Nu7441 and Mirin on A549 and NHDF‐A.
2.2.1.9 Immunostaining of H2AX foci
Immunostaining of cellular proteins involved in the DNA damage repair with fluorescent‐labeled antibodies is the most frequently used method to study DSB induction and repair. Target proteins used in immunostaining include 53BP1 (53 binding protein 1), ATM (Ataxia telangiectasia mutated) and H2AX (phosphorylated histone H2AX). Especially immunostaining of the phosphorylated histone variant H2AX is often used as a sensitive marker for DSB induction and repair. H2AX becomes phosphorylated at its Serine 139 (Ser139) residue after DSB induction to form H2AX. Responsible for this modification are kinases of the phosphatidylinositol‐3 kinase (PI3K) family, namely ATM, ATR and DNA‐PKcs. The phosphorylated H2AX localizes to the DSB site and forms microscopically‐ visible foci in a range of 1‐2 mega base pairs around the DSB site. Upon DSB repair, the phosphate residue is removed from H2AX and the foci dissociate. Hence, analyzing the dynamics in H2AX phosphorylation and dephosphorylation is often used to study the kinetics of DSB induction and repair.
For H2AX immunostaining, cells were seeded on coverslips suitable for cell cultivation in 6‐Well plates for 24h prior to DSB induction by irradiation or chemotherapy. Subsequently, cells were washed twice with 500µl 1xDPBS per well. Then, the cells were fixed with 500µl 4% paraformaldehyde for 10min at RT, rinsed twice with 500µl 1xDPBS for 5min each and subsequently permeabilized with 0.5% Triton X‐100 in 1xDPBS at RT for 7min. To remove excess Triton X‐100, the cells were rinsed twice with 500µl 1xDPBS for 5min each. Next, unspecific binding sites were blocked with 500µl 1% BSA in DPBST for 2h at RT, and H2AX foci were stained with 200µl H2AX‐specific antibody labeled with Alexa Fluor 647 (100ng/ml) in 1% BSA‐DPBST solution at 4ºC overnight on a horizontal shaker. As a negative control for immunostaining, treated cells were stained with IgG‐Alexa Fluor 647 antibody. After incubation, cells were washed twice with 500µl DPBST for 5min each. Finally, one drop of vectashield containing DAPI was added to each coverslip and the slips were viewed upside down under confocal microscope. For counting of H2AX, the Fiji platform was used [74].
2.2.1.10 Inhibition of NHEJ‐repair activity by Nu7441 and Mirin
Blocking DNA repair activity with chemical compounds increases radio‐ and chemosensitivity and can be used to study DNA repair mechanisms [75]. Here, NHEJ‐repair activity was blocked by Nu7441 (kindly provided by Dr. Friederike Herbst) and mirin. Nu7441 inhibits DNA‐PK activity and thus blocks classical NHEJ‐repair pathway. The MRN complex inhibitor mirin blocks the Mre11‐mediated activation of ATM following DNA damage, thus inhibiting end‐resection, a key step in alternative NHEJ [75]. Either Nu7441 or mirin or both together were added to IDLV‐transduced NHDF‐A at LD10 3h prior to irradiation. The cell culture medium containing the NHEJ inhibitors was removed from the cells 6h after irradiation. Cells were pooled into triplicates (3x105 cells each) 30 MATERIALS AND METHODS
and expanded until constant EGFP expression was observed. At passages 4, 6, and 8 (end point) one third of the cell populations was subjected to DNA isolation in order to analyze the clonal dynamics during expansion.
2.2.1.11 Irradiation of cells
Irradiation was performed in collaboration with the Max‐Eder‐Junior Research Group Translational Radiation Oncology at the National Center for Tumor Diseases (NCT) and German Cancer Research Center (DKFZ). Prior to irradiation, 6×106 cells were transduced with IDLV at MOI3 in 15cm cell culture dishes for 48h. Cells were subsequently transferred into 96‐Well plates at a density of 1x105 cells/well (total of 1x106 cells per radiation dose) 12h before irradiation. In order to induce DSB and DNA repair activity, cancer cell lines were irradiated with various doses of photon (X‐rays) or hadron (protons, carbons) radiation beams ranging from 0.125Gy to 10Gy. Primary human dermal fibroblasts (NHDF‐A) were irradiated with X‐rays only. Photon irradiation was delivered by a XRAD320 X‐ray device (Precision X‐Ray, CT) at 320 KeV, while proton and carbon beam therapy was performed at the Heidelberg Ion Therapy Center (HIT) with the horizontal beamline using a raster scanning technique [76]. For the detection of induced and repaired DSB sites within 24h after irradiation, 1x106 cells were detached at 1h, 4h, 8h, and 24h after irradiation and subjected to DNA isolation, LAM‐PCR and deep sequencing. In order to analyze the DSB distribution in expandable cell populations, 3x105 cells were pooled into triplicates in 6‐Well plates immediately after irradiation and expanded in cell culture until constant EGFP expression levels from stably integrated IDLV were observed. The cells were then subjected to DNA isolation and LAM‐PCR.
2.2.1.12 Inhibition of Topoisomerase 2 in mammalian cells with doxorubicin and etoposide
Doxorubicin and etoposide belong to the drug class of topoisomerase 2 poisons. Both chemicals inhibit the active TOP2 dimer from religation of the G segment. The stalled TOP2 is therefore removed by proteolytic enzymes and nucleases from the DNA, thereby exposing the DNA double strand break to the repair machinery.
Prior to TOP2 poisoning, 3x106 NHDF‐A and Hela cells were transduced with IDLV (MOI3) in 15cm cell culture dishes for 48h and subsequently seeded in 96‐Well plates at a density of 1x105 cells per well. Then, NHDF‐A and Hela cells were treated with doxorubicin or etoposide for seven subsequent days. The DMEM cell culture medium containing the TOP2 poisons was replaced with fresh poison containing cell culture medium on days 3 and 5. Following treatment, 3x105 cells were pooled into triplicates in 6‐Well plates and expanded in cell culture until constant EGFP expression levels from stably‐integrated IDLV were observed. NHDF‐A cells were subjected to DNA isolation and LAM‐PCR.
2.2.1.13 Preparation for FACS
Fluorescent Activated Cell Sorting (FACS) is a special method of flow cytometry. It allows cells to be sorted based upon specific physical properties such as light scattering or fluorescent characteristics of the cell. A single‐cell suspension passes a laser with a distinct wave length and the emission is recorded. The cells can be either stained by fluorescent‐tagged antibodies binding to cell‐type specific target molecules or the cells express a fluorescent protein themselves. In addition to fluorescence intensity, the cell size and granularity can be 31 MATERIALS AND METHODS
determined. In this thesis, enhanced green fluorescent protein (EGFP)‐expression from lentiviral vectors was measured in order to determine transduction efficiency and follow integration of IDLV during cultivation. Additionally, flow cytometry was used to determine cell cycle phase in synchronized NHDF‐A. EGFP is excited at a wave length of 488nm and has a single major emission peak at 509nm.
For flow cytometry measurement of EGFP, the cells were washed with 1xDPBS, detached from the cell culture dish with 0.05% trypsin at 37°C for 7min and transferred to a fresh polypropylene tube. Cells were centrifuged at 800 rpm for 5min at RT. Subsequently, the cells were washed once with 1xDPBS, centrifuged, and dead cells stained with propidium iodide in 2% BSA‐DPBS solution. Then, cells were centrifuged at 800 rpm for 5min at RT, and resuspended in 500µl 2% BSA‐DPBS. Cells were stored on ice until FACS analysis.
For determining the cell cycle phase of NHDF‐A, the cells were washed once with 1xDPBS, detached from the cell culture dish with 0.05% trypsin at 37°C for 7min and transferred to a fresh propylene tube. Cells were centrifuged at 1500xg for 5min at RT. Subsequently, the cells were washed once with 1xDPBS, centrifuged, and fixed by adding 1ml ice‐cold 70% ethanol dropwise 0.3ml cell suspension in DPBS. Cells were left on ice for 2h. Subsequently, cells were centrifuged, the supernatant discarded and cells washed twice with 1ml DPBS. Since PI does not exclusively bind to nuclear DNA, but also to double‐stranded RNA, the RNA needed to be degraded. Therefore, NHDF‐A were incubated with 75µl ribonuclease A (Rnase A) (10mg/ml) at 37°C for 30min. The cells were subsequently mixed with 1ml DPBS, centrifuged and stained with propidium iodide in 2% FCS‐DPBS solution on ice for 30min. Then, cells were centrifuged at 1200xg at RT for 5min, and resuspended in 500µl 2% FCS‐DPBS. Cells were stored on ice until flow cytomtric analysis. Flow cytometry was performed on a LSRII from BD Biosciences.
2.2.1.14 Synchronization of NHDF‐A in G1/G0, S and G2/M phase of the cell cycle
Cell synchronization is a process by which cells in different cell cycle phases are brought to the same phase. Several different strategies including addition of chemical substances or depletion of factors in the cell culture medium are used to achieve synchronization. Here, NHDF‐A were synchronized in G1/G0, S and G2/M phase of the cell cycle by serum starvation, double thymidine block and nocodazole treatment, respectively.
In order to synchronize cells, the cells need to be seeded at a confluency of 30‐50% to enable cell division and prevent contact inhibition. Elimination of serum from the culture medium results in the accumulation of cells in G1/G0 phase. Therefore, DMEM supplemented with 10% FCS was removed from the cells and replaced with DMEM containing 1% FCS for 72h. In order to synchronize cells in S phase, DNA synthesis can be blocked by using the inhibitor thymidine. The cell culture medium was removed from the NHDF‐A and replaced with DMEM supplemented with 10% FCS and 2mM thymidine for 48h. This first block was followed by a 16h incubation period with DMEM supplemented with 10% FCS. Then, a second thymidine block was induced by replacing the medium with DMEM supplemented with 10% FCS and 2mM thymidine for 72h. Nocodazole interferes with the polymerization of microtubules at the end of G2 phase and beginning of mitosis, thereby inhibiting entry of cells into mitosis. Here, NHDF‐A were synchronized in G2/M by addition of nocodazole at a concentration of 100ng/ml to DMEM supplemented with 10% FCS. Efficiency of synchronization was analyzed by flow cytometry.
32 MATERIALS AND METHODS
2.2.2 Molecular Biology Methods
2.2.2.1 Isolation of genomic DNA from cultivated cells
For the isolation of genomic DNA, the High Pure PCR Template Preparation Kit from Roche was used according
to the manufacture’s protocol with a single modification. The DNA was eluted twice with 50µl dH2O each into a single 1.5ml reaction tube. The DNA concentration was measured at the Nanodrop‐1000 and stored at ‐20°C.
2.2.2.2 Determining the DNA concentration
The concentration of the genomic DNA was measured using either the Nanodrop‐1000 or Qubit. For Nanodrop DNA analysis, 1.5µl DNA were transferred to the Nanodrop‐1000 machine. The DNA concentration was determined spectrometrically at a wavelength of 260nm. For Qubit measurements, the Qubit dsDNA HS assay kit was used according to manufacturer’s protocol.
2.2.2.3 Polymerase Chain Reaction (PCR)
The Polymerase Chain Reaction (PCR) enables the exponential amplification of a target DNA for cloning, induction of mutations or sequencing [77]. The PCR can be subdivided into three basic steps, which are repeated for 20‐50 cycles. Each cycle starts with the denaturation of the double‐stranded DNA by heating the DNA to 95°C. Prior to the first PCR cycle, the DNA is usually heated for a longer time to guarantee that the two DNA strands are separated completely before the cycling begins. Subsequently, the DNA is cooled to allow primers to anneal to the target sites. The exact temperature for primer annealing is calculated based on the primer length and its base composition. Usually, annealing temperatures are about 60°C. If the annealing temperature is too low, primers can bind to unrelated DNA sites and result in the amplification of unspecific sequences. In the final step, called elongation phase, the temperature is increased up to 68°C‐72°C and the DNA is amplified by the DNA polymerase starting from the 3’ end of the primers. During elongation, the primer becomes integrated into the amplified DNA (amplicon). The repetition of the three phases (denaturation, annealing, elongation) results in the exponential amplification of the target DNA sequence.
Reagent Amount
H2O bidest Ad 50µl 10x Taq DNA Polymerase Buffer 5µl dNTP mix (dATP, dGTP, dCTP, dTTP) (10mM each) 1µl Sense Primer (10pmol/µl) 20pmol Antisense Primer (10pmol/µl) 20pmol Taq DNA Polymerase (1U/µl) 2µl
For amplification, the following PCR cycling program was used for a Thermus aquaticus (Taq) DNA polymerase:
33 MATERIALS AND METHODS
Description Temperature Time Repetitions Initial Denaturation 95°C 5min Denaturation 95°C 30sec Annealing 60°C 30sec 30 cycles Elongation 72°C 1min Final Elongation 72°C 5min Cooling 4°C
2.2.2.4 DNA agarose electrophoresis
Gelectrophoresis is a frequently used method for the separation and visualization of DNA. The DNA fragments are separated according to their molecular mass in an agarose gel when an electric charge is applied. Since the DNA is negatively charged, it is pulled towards the anode. According to the size of the DNA fragments to be separated, 0.8‐2% agarose concentrations are used.
First, the agarose was dissolved in 100ml 1xTris‐borat‐EDTA (TBE) buffer and subsequently heated in a microwave until the agarose powder was dissolved completely. Then, one drop of ethidiumbromide solution was added to the liquid agarose gel, and the gel was poured into a gel chamber. Upon solidification, the gel chamber was placed into an electrophoresis chamber with 1xTBE buffer. Then, 8µl DNA were mixed with 2µl 5x loading buffer and transferred into the wells of the gel. Finally, the DNA fragments were separated at 130V for 45 minutes. Since the ethidiumbromide present in the gel binds to the DNA and fluorescates in UV light, the DNA fragments was visualized upon UV light exposure. The DNA fragment sizes were determined by comparing the DNA bands to a marker of known size.
2.2.2.5 DNA isolation from agarose gel
For isolating DNA after gel electrophoresis from agarose gels, a piece of the gel containing the DNA fragment of interest was excised and transferred into a 1.5ml reaction tube. Subsequently, the DNA was isolated by using the QIAqucik Gel Extraction Kit from Qiagen according to manufacturer’s protocol.
2.2.2.6 Absolute quantitative real‐time PCR (q‐RT‐PCR)
Real‐time quantitative PCR is special type of PCR, which enables amplification and quantification of DNA simultaneously. Quantification is performed by measuring the incorporation of a fluorescent dye into newly synthesized double stranded DNA during amplification. Since the amount of DNA increases during each PCR cycles and thus the fluorescence, the fluorescence intensity can used to calculate the amount of DNA in each reaction. Here, the fluorescent dye used was SYBR green I, which is measured after each PCR cycle. SYBR Green is excited 488nm and it emits green light at a wavelength of 522nm. By comparing the fluorescent levels of the DNA samples to a standard dilution series of known concentration the DNA concentration in the sample can be calculated. To ensure that different samples can be compared to each other, the expression of the target gene is normalized to the expression of a reference gene present in every sample. Here, the target gene was the EGFP locus in IDLV, and the genomic myoglobin (MYO) gene locus was used as a reference. 34 MATERIALS AND METHODS
Prior to each absolute q‐RT‐PCR, qPCR standards for the target and reference locus were synthesized. Therefore, the EGFP locus and the MYO locus were amplified by PCR using an annealing temperature of 60°C. Subsequently, the PCR fragments were analyzed by gel electrophoresis and isolated from the agarose gel. Then, the DNA concentrations were determined at the Nanodrop‐1000, and the amplicon copy number (n) calculated using the following formula:
6.0223 10 10 650
For the EGFP (326bp) and MYO (339bp) standards, 1x108 DNA copies were inserted into the q‐RT‐PCR reaction, which is equal to 0.0352ng and 0.0367ng, respectively. Subsequently, the EGFP and MYO standard dilutions series ranging from 106 to 100 copies were prepared in 5µl ms2RNA solution per reaction. For each DNA sample obtained from IDLV‐transduced and irradiated cells, 30ng genomic DNA diluted in 5µl ms2RNA solution were inserted for the amplification of the MYO genomic locus and the EGFP gene encoded in the IDLV. Subsequently, the following mastermix was prepared for each primer pair:
Reagent Amount
H2O bidest 4µl Primer 1 (10mM) 0.5µl Primer 2 (10mM) 0.5µl 2x SYBR Green I 10µl
For each standard dilution and DNA sample, triplicate reactions were prepared. The q‐RT‐PCR reactions were prepared in 96‐Well plates specifically designed for q‐RT‐PCR. After the addition of the mastermix to each well, the 96‐Well plate was sealed and centrifuged for 3min at 800xg. The qPCR was performed in a Lightcycler 480 using the following program:
Description Target Temperature Time Ramp Rate Repetitions Pre‐Incubation 95°C 15min 4.4°C/sec 95°C 10sec 2.2°C/sec Amplification 58°C 5sec 4.4°C/sec 40 cycles 72°C 25sec 4.4°C/sec 95°C 5sec 4.4°C/sec Melting Curve 65°C 1min 2.2°C/sec 97°C 0.11°C/sec Cooling 40°C 30sec 2.2°C/sec
2.2.2.7 Linear‐Amplification Mediated Polymerase Chain Reaction (LAM‐PCR)
To detect and analyze the integration site distribution of retroviruses in the genome, linear‐amplification mediated polymerase chain reaction (LAM‐PCR) was developed [78, 79]. In the first step of LAM‐PCR, the retroviral vector‐genome junction is amplified by a linear PCR step. Biotinylated primers bind to the LTR sequence of the IDLV and are subsequently elongated by the Taq polymerase. The resulting single‐stranded amplicons are immobilized on paramagnetic beads which carry streptavidin on their surface, resulting in the 35 MATERIALS AND METHODS
separation of the amplified sequences from the remaining genomic DNA. The following steps are performed with the solid phase. The single‐stranded DNA is converted into double‐stranded DNA using the bacterial Klenow Polymerase. Afterwards, the immobilized dsDNA is digested by a restriction enzyme that cuts in the genomic part of the amplicon to yield a single‐stranded DNA overhang. This overhang is used to ligate a DNA linker of known sequence called the linker cassette to the amplicon. The non‐biotinylated DNA strand is subsequently separated by denaturation and amplified by two rounds of nested PCR with vector‐ and linker‐specific DNA primers. The amplified DNA is then subjected to pyrosequencing (Figure 9).
Figure 9: Schematic of 5’ LAM‐PCR for the amplification of retroviral integration sites. (a) Linear amplification of the vector‐ genome junction using biotinylated primers which bind in the U3 region of the LTR. (b) The amplicons from the linear amplification are immobilized to paramagnetic beads. (c) The single‐stranded amplicon is converted into double‐stranded DNA using Klenow polymerase from E.coli. (d) Enzymatic restriction digest of the genomic part of the amplicons to create single‐stranded overhangs. (e) Ligation of DNA linker sequences to the ssDNA overhangs. (f) Alkaline denaturation detaches the non‐biotinylated DNA, which is subsequently used for two rounds of exponential amplification (g). Since LTR are present at both ends of the retroviral genome, two fragments are amplified: vector‐genome junction and an internal vector sequence. B: biotinylated primer; LTR: long terminal repeat; LTR I: oligonucleotide for the linear amplification; LTR II: oligonucleotide for the first exponential PCR; LTR II: oligonucleotide for the second exponential PCR; LCI: DNA linker‐specific oligonucleotide for first exponential PCR; LCII: DNA linker‐specific oligonucleotide for first exponential PCR. Image taken from [78]
Creation of linker cassettes
Two single‐stranded DNA oligos that are complementary to each other can be hybridized to form a double‐ stranded DNA molecule. This dsDNA linker cassette is used in the ligation reaction. According to the restriction enzyme used during LAM‐PCR, the sequence of second linker cassette is chosen and used for hybridization.
Basically, the two oligonucleotides are mixed in a hybridization buffer:
Reagent Amount LC 1 (100mM) 40µl LC 3 (100mM) 40µl
MgCl2 (100mM) 10µl Tris‐HCl pH7.5 (250mM) 110µl
36 MATERIALS AND METHODS
The solution was prepared in a 1.5ml reaction tube and heated to 95°C for 5min in a heating block. After the incubation, the heating block was switched off and the tube was cooled in the heating block over night.
Subsequently, the solution was filled up to a total volume of 500µl with H2O bidest., transferred to a Microcon YM‐30 centrifugation filter and centrifuged at 14,000 x g for 10min. The linker cassette was then eluted from the inverted filter tube into a fresh 1.5ml reaction tube at 1,000 x g for 2min. Finally, the volume of the oligonucleotide solution was adjusted to 80µl, and stored at ‐20°C in 10µl aliquots.
Linear PCR
For linear amplification, 500ng or 1µg of genomic DNA was inserted. To increase the total number of IDLV integration sites, triplicates for each sample were inserted into LAM‐PCR. Following the first 50 cycles of linear amplification, 1µl Taq DNA polymerase (5U) was added to each sample, and amplification was repeated for another 50 cycles. IDLV transduced samples were amplified with the biotinylated primers SK‐LTR 1bio and SK‐LTR
3bio (0.167µM each). As controls for LAM‐PCR, non‐transduced human genomic DNA from Roche and H2O bidest were used.
Reagent Amount
H2O bidest Ad 50µl 10x Taq DNA Polymerase Buffer 5µl dNTP mix (dATP, dGTP, dCTP, dTTP) (10mM each) 1µl Primer 1 (0.167µM) 0.25µl Primer 2 (0.167µM) 0.25µl Taq DNA Polymerase (5U/µl) 0.25µl DNA 500ng/1µg
Description Temperature Time Repetitions Initial Denaturation 95°C 2min Denaturation 95°C 45sec Annealing 60°C 45sec 2x50 cycles Elongation 72°C 1min Final Elongation 72°C 5min Cooling 4°C
Magnetic Capture of amplicons
For each PCR reaction, 20µl magnetic beads were prepared. The magnetic beads were transferred into a fresh 1.5ml reaction tube and placed in a magnetic particle concentrator (MPC) for 1min. The storage solution was discarded and the beads resuspended in 40µl 0.1% BSA‐DPBS solution per reaction. Again, the tube was placed in the MPC and the supernatant discarded. The washing step was subsequently repeated. Afterwards, the beads were washed with 20µl of the 3M LiCl solution per reaction. Finally, the beads were resuspended in 50µl 6M LiCl solution per reaction and mixed with 50µl linear PCR product in the PCR tube. To allow biotinylated DNA amplicons to bind to the streptavidin‐coated paramagnetic beads, the tubes were placed in a horizontal shaker at 300 rpm and RT over night.
Hexanucleotide priming
During hexanucleotide priming, the ssDNA amplicons from the linear amplification were converted into dsDNA by the bacterial Klenow polymerase. The reaction mastermix was prepared on ice as followed: 37 MATERIALS AND METHODS
Reagent Amount per reaction
H2O bidest 16.5µl 10x Hexanucleotide Mix 2µl dNTP mix (dATP, dGTP, dCTP, dTTP) (10mM each) 0.5µl Klenow DNA polymerase 1µl
To each PCR tube, 80µl H2O bidest was added and the beads transferred into a 96‐Well plate. Subsequently, the beads were exposed to a MPC for 1min and the supernatant discarded. Washing was repeated once with 100µl
H2O bidest and the 96‐Well plate was put in a cooling rack. Finally, 20µl hexanucleotide mastermix was added to each well and incubation was performed at 37°C for 1h.
Restriction digest of amplicons
The DNA‐paramagnetic particle complexes were washed once with 80µl and once with 100µl H2O bidest. The bead solution was resuspended in 20µl of the restriction digest mastermix.
Reagent Amount per reaction
H2O bidest 8.7µl 10x CutSmart buffer 1µl Restriction Enzyme (5U/µl) 0.3µl
Restriction enzymes, which do not have a recognition site in the LTR of IDLV were chosen for LAM‐PCR. Here, MseI and MluCI/Tsp509I were used. The restriction mixture was incubated in a PCR cycler at 37°C (MseI, MluCI) or 64°C (Tsp509I) for 2h.
Ligation of linker cassette
To each restriction digest mixture, 80µl H2O bidest were added and the 96‐Well plate placed on a MPC for 1min.
The supernatant was discarded and the washing step repeated with 100µl H2O bidest The DNA‐particle mix was resuspended in 10ul ligation mastermix.
Reagent Amount per reaction
H2O bidest 5µl 10x Fast Link Ligation Buffer 1µl ATP (10µM) 1µl Fast Link DNA ligase (2U/µl) 1µl Linker cassette 2µl
The ligation was performed at RT for 5min. To stop ligation reaction, the 96‐Well plate was placed on ice. For Illumina MiSeq pyrosequencing, linker cassettes with unique DNA recognition sequences, so called barcodes, for each sample were used in the ligation reaction.
Denaturation
The total volume in each well was filled up to 100µl with 90µl H2O bidest. Then, the 96‐Well plate was placed on
a MPC, the supernatant discarded and washing repeated with 100µl H2O bidest. Subsequently, the beads were resuspended in 5µl 0.1N NaOH solution and incubated at 300rpm on a horizontal shaker and at RT for 10min. 38 MATERIALS AND METHODS
Again, the 96‐Well plate was placed on a MPC and the supernatant transferred into fresh 0.5ml reaction tubes. The DNA was either stored at ‐20°C or used immediately for exponential amplification.
Nested PCR for exponential amplification of vector‐genome junctions
The exponential PCR was performed using nested primers that bind specifically to the LTR and linker cassette sequence. For the first nested PCR, 1µl of linear PCR product and the biotinylated primers SK‐LTR 4bio (3’ LAM‐ PCR) as well the linker cassette specific primer LCI were inserted.
Reagent Amount per reaction
H2O bidest Ad 25µl 10x Taq DNA Polymerase Buffer 2.5µl dNTP mix (dATP, dGTP, dCTP, dTTP) (10mM each) 0.5µl Primer 1 (16.7µM) 0.125µl LCI (16.7µM) 0.125µl Taq DNA Polymerase (5U/µl) 0.25µl DNA 1µl
Description Temperature Time Repetitions Initial Denaturation 95°C 2min Denaturation 95°C 45sec Annealing 60°C 45sec 35 cycles Elongation 72°C 1min Final Elongation 72°C 5min Cooling 4°C
For capturing of the biotinylated PCR products, 20µl magnetic beads per PCR reaction were prepared as described (see magnetic capture of amplicons). The beads were resuspended in 25µl 6M LiCl solution per reaction and mixed with 25µl PCR product in the PCR tube. To allow biotinylated DNA amplicons to bind to the streptavidin‐coated paramagnetic beads, the tubes were placed in a horizontal shaker at 300 rpm for 2h.
Afterwards, 80µl H2O bidest was added to each PCR reaction tube and the beads transferred into a new 96‐Well plate, which was subsequently placed on a MPC. The supernatant was discarded and the washing step repeated
twice with 100µl H2O bidest. Next, beads were resuspended in 5µl 0.1N NaOH, and the plate was incubated at 300rpm in a horizontal shaker at RT for 10min. The plate was placed on a MPC and the supernatant transferred into a fresh 0.5ml reaction tube. For the second exponential PCR, 1µl denaturation product was inserted. The linker‐specific primer used in this PCR was LCII, and the vector‐specific primer was either SK‐LTR 5 (3’ LAM‐PCR).
Reagent Amount per reaction
H2O bidest Ad 50µl 10x Taq DNA Polymerase Buffer 5µl dNTP mix (dATP, dGTP, dCTP, dTTP) (10mM each) 1µl Primer 1 (16.7µM) 0.25µl LCII (16.7µM) 0.25µl Taq DNA Polymerase (5U/µl) 0.5µl DNA 1µl
39 MATERIALS AND METHODS
Description Temperature Time Repetitions Initial Denaturation 95°C 2min Denaturation 95°C 45sec Annealing 60°C 45sec 35 cycles Elongation 72°C 1min Final Elongation 72°C 5min Cooling 4°C
To analyze the LAM‐PCR products, 2µl of the amplicons were mixed with 8µl loading dye and loaded onto a 2% agarose gel.
2.2.2.8 Cleaning‐Up of PCR products using AMPure XP beads
In order to attach sequencing adaptors to PCR amplicons or to subject samples to pyrosequencing, the LAM‐PCR products were cleaned up. Here, AMPure XP beads from Beckman Coulter were used according to manufacturer’s protocol. For the attachment of sequencing adaptors, the beads‐to‐PCR product ratio was 1.1:1. For pyrosequencing, the ratio of beads‐to‐PCR product was changed to 0.65:1.
2.2.2.9 Fusionprimer‐PCR
Fusionprimer‐PCR was performed to attach Roche/454‐ or Illumina‐specific sequencing adapters to the LAM‐PCR amplicons. For Roche/454 sequencing, an adaptor sequence was attached to the linker sequence in order to enable binding of the PCR amplicons to specific beads during sequencing. At the vector sequence, an adaptor was attached in order to enable sequencing. For Illumina sequencing, two primers were attached to the vector‐ and linker sequence to facilitate binding of the amplicons to an array surface during sequencing. The vector‐ specific adaptor contained an additional 6 or 10bp long unique recognition sequence, called barcode, which enables massive parallel sequencing and later sorting of the sequences. For 454 and Illumina sequencing, 50ng cleaned‐up LAM‐PCR products were inserted into PCR. Each Fusion‐PCR primer consists of the specific sequencing‐adaptor and a vector‐ or linker‐specific sequence.
Reagent Amount per reaction
H2O bidest Ad 50µl 10x Taq DNA Polymerase Buffer 5µl dNTP mix (dATP, dGTP, dCTP, dTTP) (10mM each) 1µl Fusionprimer forward (5µM) 0.5µl Fusionprimer reverse (5µM) 0.5µl Taq DNA Polymerase (5U/µl) 0.5µl LAM‐PCR amplicon 50ng
40 MATERIALS AND METHODS
Description Temperature Time Repetitions Initial Denaturation 95°C 2min Denaturation 95°C 45sec Annealing 60°C 45sec 15 cycles Elongation 72°C 1min Final Elongation 72°C 5min Cooling 4°C
The PCR products were subjected to gel electrophoresis, cleaned up, and measured spectrometrically using the Nanodrop‐1000.
2.2.2.10 SureSelect Target Enrichment for Illumina Multiplexed Sequencing
The SureSelect Target Enrichment protocol is used to enrich target regions of the genome from unrelated sequences and prepare the samples for Illumina sequencing. Prior to enrichment, pre‐capture libraries were prepared that target specific sequences. These target sequences among others include SIN lentiviral vectors. The protocol is divided into sample preparation, hybridization, sample processing for multiplexed sequencing, and sequencing (Figure 10).
Figure 10: SureSelecet Target enrichment workflow for Illumina sequencing. In the first step, the genomic DNA is sheared into fragments of 150‐250bp size. Each sample is subsequently indexed, amplified and pooled with other indexed libraries. Target sequences are then enriched using a SureSelect Pre‐Capture Library, captured and amplified by PCR. The indexed library is finally subjected to Illumina HiSeq sequencing. Image taken from Agilent Technologies manual (SureSelect XT2 Target Enrichment System for Illumina Multiplexed Sequencing, Version D.0, September 2013).
41 MATERIALS AND METHODS
Sample Preparation
Step 1: Shear genomic DNA
A total of 1µg genomic DNA of IDLV‐transduced, irradiated and expanded NHDF‐A was diluted in 50µl H2O bidest. The solution was transferred into a Covaris microtube and subsequently sheared using the following settings: Duty Factor: 10%, Intensity Level: 5, Cycles per Burst: 200, Treatment Type: 120sec, yielding DNA fragments with an average size of 250bp. Then, the DNA fragments were transferred into fresh 0.5ml reaction tubes and stored at ‐20°C until further usage.
Step 2: Assess quality of sheared DNA (skipped)
Step 3: Repair the ends
In order to repair the DNA ends, 50µl SureSelect End‐Repair Mastermix was added to each sheared DNA sample and mixed well by pipetting the solution up and down. The mixture was incubated at 20°C in a thermal cycler without using the heated lid. Finally, the mixture was cooled down and kept at 4°C.
Step 4: Purify the samples using AMPure XP beads
Prior to purification, the AMPure XP beads were warmed at RT for 30 minutes. For each sample, 500µl 80% ethanol was prepared. Then, 160µl beads were mixed with 100µl end‐repaired DNA (beads‐to‐DNA ratio: 1.6:1) and mixed by pipetting up and down ten times, followed by 5min incubation at RT. Then, the 0.5ml reaction tubes were placed into a MPC for 4 minutes and the supernatant removed. While washing each sample twice with 200µl 80% ethanol, the reaction tubes were kept in the magnetic stand. The residual ethanol was removed,
the tubes removed from the MPC, 22µl nuclease‐free H2O added to each sample and incubated for 2min at RT before placing the tubes back into the MPC. The DNA‐containing solutions were transferred to fresh 0.5ml reaction tubes.
Step 5: Adenylate 3’ ends of DNA fragments
To add a single adenine nucleotide to the 3’ end of the DNA fragments, 20µl dA‐Tailing Master Mix was added to each sample, mixed and incubated in thermal cycler at 37°C for 30 minutes without using the heated lid. The mixture was cooled down and kept at 4°C.
Step 6: Ligate the pre‐capture indexing adaptors
In this step, DNA adaptors carrying unique DNA barcodes, called index, were ligated to each DNA sample to generate an indexed library. To each A‐tailed DNA sample, 5µl Ligation Master Mix and 5µl of the appropriate Pre‐Capture Index Solution were added. The mixtures were vortexed for 5 seconds, spinned briefly in a table‐top centrifuge, and incubated in thermal cycler at 20°C for 15min without using the heated lid.
Step 7: Purify the indexed DNA using AMPure XP beads
Purification was performed as described above (step 4) with two changes. To each index sample, 36µl beads
were added (Beads‐to‐DNA ratio: 1.2:1), and the purified DNA was resuspended in 50µl nuclease‐free H2O.
Step 8: Amplify the indexed library 42 MATERIALS AND METHODS
For amplification, the following Pre‐capture PCR reaction mix was prepared on ice. For each indexed sample, 1µl XT2 Primer Mix and 25µl Herculase II PCR Master Mix were mixed and combined with 24µl of each indexed library in a fresh 96‐Well PCR plate. The following program was run in a PCR thermal cycler:
Temperature Time Repetitions 98°C 2min 98°C 30sec 60°C 30sec 5 Cycles 72°C 1min 72°C 10min 4°C hold
Step 9: Purify the amplified library with AMPure XP beads
Purification was performed as described above (step 8) with a single change ion protocol: to each amplified library, 50µl beads were added (Beads‐to‐DNA ratio: 1:1).
Step 10: Assess quality with the 2100 Bioanalyzer DNA 1000 Assay
To assess the quality of each indexed library, 1µl of each sample and 1µl D1000 Ladder were mixed with 3µl D1000 Sample Buffer each in a tube strip, vortexed for 3 sec and placed in a 2200 TapeStation. Finally, gel electrophoresis was run and quality of amplified libraries analyzed.
Hybridization
Step 1: Pool indexed DNA samples for hybridization
For each hybridization reaction, a total of 2000ng indexed DNA library was required. Here, 19 amplified libraries were combined (111ng of each library) in a single 1.5ml reaction tube. Subsequently, the samples were completely dried in a vacuum concentrator at 37°C for 3h. Then, the DNA was resuspended in 7µl TE buffer (PCR purification kit, Qiagen) and dissolved for 4h at 300rpm and RT in a horizontal shaker.
Step 2: Hybridize DNA library pools to the SureSelect Capture Library
To each pooled genomic DNA (gDNA) library, 9µl SureSelect Blocking Mix was added and the solution transferred into a fresh PCR reaction tube. The tube was placed in PCR cycler at 95°C for 5min and then held at 65°C. In the meantime, the Capture Library/RNase Block Mix was prepared by mixing 1µl SureSelect Capture Library and
0.5µl SureSelect RNase Block with 5.5µl H2O bidest. Subsequently, 37µl SureSelect hybridization buffer was added to the 7µl Capture Library/RNase Block Mix. While maintaining the gDNA pool at 65°C, 44µl Capture Library Mix was added and mixed by pippetting up and down. The hybridization mixture was kept at 65°C for 67h with the heated lid at 99°C.
Step 3: Prepare streptavidin magnetic beads for DNA hybrid capture
First, SureSelect Wash buffer #2 was pre‐warmed at 65°C in a water bath. Second, 50µl Dynabeads MyOne Streptavidin T1 were transferred to a fresh 0.5ml reaction tube and washed by adding 200µl SureSelect Binding buffer. The tube containing the beads was vortexed for 5sec, and the supernatant removed from the beads on a 43 MATERIALS AND METHODS
MPC. This washing step was repeated for a total of three washes. Finally, beads were resuspended in 200µl SureSelect Binding buffer.
Step 4: Capture the hybridized DNA using streptavidin beads
The hybridization mixture was kept at 65°C while the entire hybridization mixture was transferred into the tube containing the 200µl washed Dynabeads from step 3. The solution was mixed, and incubated on a horizontal shaker at RT and 300rpm for 30min. The tube was placed in a MPC and the supernatant removed. The beads were resuspended in 200µl SureSelect Wash buffer #1, mixed, and the tube placed back into the MPC. The supernatant was subsequently removed and the beads washed with 200µl pre‐warmed SureSelect Wash buffer #2. The beads were resuspended, incubated at 65°C for 5min in a heating block, and the tube placed into a MPC. The supernatant was discarded and washing with buffer #2 was repeated for a total of 6 washes. Finally, the
beads were mixed with 30µl nuclease‐free H2O and stored at 4°C.
Post‐Capture Sample Processing for Multiplexed Sequencing
Step 1: Amplify the captured libraries
For amplification of the captured library, the following PCR reaction mix was prepared on ice. For each library,
25µl Herculase II Master Mix, 1µl XT2 Primer Mix and 9µl nuclease‐free H2O were mixed and added to 15µl of the captured library pool bead suspension in a PCR reaction tube. The following program was run in a PCR thermal cycler:
Temperature Time Repetitions 98°C 2 min 98°C 30 sec 60°C 30 sec 14 cycles 72°C 1 min 72°C 10 min 4°C hold
Step 2: Purify the amplified captured libraries using AMPure XP beads
Purification was performed as described above (Sample Preparation, step 4) with two changes: To each amplified and captured library, 80µl beads were added (Beads‐to‐DNA ratio: 1.6:1). The beads were
resuspended in 15µl H2O bidest, the tube was placed in a MPC and the supernatant transferred to a fresh reaction tube.
Step 3: Assess quality with the 2100 Bioanalyzer High Sensitivity DNA assay
For assessing the quality of the amplified and captured library, 1µl of each library and 1µl D1000 Ladder were mixed with 3µl D1000 Sample Buffer each in a tube strip, vortexed for 3sec and placed in a 2200 TapeStation. Finally, gel electrophoresis was run and quality of amplified libraries was analyzed.
Step 4: Prepare samples for multiplexed sequencing 44 MATERIALS AND METHODS
To sequence the DNA using HiSeq 2000, 3µl of the DNA solution was diluted 1:10 in H2O bidest. Sequencing was performed at the DKFZ Genomic and Proteomics Core Facility.
2.2.2.11 Tdt‐mediated labeling of DSB sites
For the direct labeling of free DNA ends the TUNEL assay (TdT‐mediated dUTP‐biotin nick end labeling) protocol was modified.
Fixation and permeabilization
In the first step, 1x106 A549 cells were washed once with 1xDPBS and centrifuged at 1,000rpm and RT for 5min. Then, the cells were fixed with 4% paraformaldehyde (PFA) at 37°C and 600rpm for 10min in a heating block. Subsequently, the fixed cells were washed with 1xDPBS before permeabilization with 0.25% Triton X‐100 solution for 20min in a heating block at 37°C and 600rpm.
dUTP labeling of free DNA ends
Residual PFA and Triton X‐100 solution were removed by washing the cells twice with H2O bidest. Centrifugation was performed at 1,000rpm for 5min at RT. In order to label free DNA ends with UTP nucleotides, the following Tdt master mix was prepared in a 1.5ml reaction tube is shown in the table below:
Reagent Amount
H2O bidest 53.5 µl 5x Tdt reaction buffer 20 µl
CoCl2 solution (25mM) 20 µl Biotin‐16‐dUTP (1mM) 5 µl dUTP (100mM) 0.5 µl Tdt (400U/µl) 1 µl
As controls for the labeling reaction, EcoRI‐ or PvuI‐digested LV902 plasmid and H2O bidest were included. The cells and control labeling reactions were incubated with the Tdt master mix at 37°C and 600rpm in a heating block for 60min. In order to block Tdt labeling activity, the 1.5ml was placed on ice and 10µl 0.2M EDTA (pH8) added. Then, the total volume was filled up to 200µl with 1xDPBS.
DNA isolation
To isolate the genomic DNA from the fixed cells, the cells were lysed with 200µl lysis buffer (4M urea, 200mM Tris‐HCl, 20mM NaCl, 200mM EDTA, pH7.4) and 40µl proteinase K (1µg/ml) for 1h at 55°C. The DNA was subsequently precipitated by adding 440µl isopropanol. Next, the DNA was centrifuged for 15min at full speed in a table top centrifuge. The supernatant was carefully removed and the DNA washed with 100µl 70% ethanol. The centrifugation step was repeated, the ethanol removed and the DNA dried at RT for 5min. Finally, the DNA was resuspended in 50µl TE buffer.
Magnetic Capture of amplicons
For each DNA sample, 20µl magnetic beads were transferred into a fresh 1.5ml reaction tube and placed in a magnetic particle concentrator (MPC) for 1min. The storage solution was discarded and the beads resuspended 45 MATERIALS AND METHODS
in 40µl 0.1% BSA‐DPBS solution per reaction. Again, the tube was placed in the MPC and the supernatant discarded. The washing step was repeated once. Afterwards, the beads were washed with 20µl of the 3M LiCl solution per reaction. Finally, the beads were resuspended in 50µl 6M LiCl solution per reaction and mixed with 50µl labeled genomic DNA in a 96‐Well plate. To allow biotinylated DNA to bind to the streptavidin‐carrying paramagnetic beads, the plate was placed in a horizontal shaker at 300 rpm and RT over night.
Restriction digest of captured DNA
The DNA‐paramagnetic particle complexes were washed once with 100µl 0.1% Tween‐20 solution and once with
100µl H2O bidest. The bead solution was resuspended in 20µl of the restriction digest mastermix.
Reagent Amount per reaction
H2O bidest 17.6µl 10x NEBuffer 1 2µl Tsp509I (5U/µl) 0.4µl
The restriction mixture was incubated in a PCR cycler at 65°C (Tsp509I) for 1h.
Ligation of linker cassettes
To block restriction enzyme activity, 80µl H2O bidest were added to each reaction, and the 96‐Well plate placed
on a MPC for 1min. The supernatant was discarded and the washing step repeated with 100µl H2O bidest. The DNA‐particle mix was resuspended in 10µl ligation mastermix.
Reagent Amount per reaction
H2O bidest 5µl 10x Fast Link Ligation Buffer 1µl ATP (10µM) 1µl Fast Link DNA ligase (2U/µl) 1µl Linker cassette 2µl
The ligation was performed at RT for 5min. To stop ligation reaction, the 96‐Well plate was placed on ice. The linker cassettes inserted into the ligation reaction were the same ones used for LAM‐PCR and had an AATT DNA overhang complementary to the free DNA overhang introduced by the restriction enzyme Tsp509I.
Nested PCR for exponential amplification of linker‐genome junctions
The total volume in each well was filled up to 100µl with 90µl H2O bidest. Then, the 96‐Well plate was placed on
a MPC, the supernatant discarded and washing repeated with 100µl H2O bidest. The exponential PCR was performed using nested PCR primers that bind specifically to the polyU tail and linker cassette sequence. For the first PCR, 2µl DNA‐beads solution and the polyA‐primer L.CDSB2‐1 as well as the linker cassette specific primer LCI were inserted.
46 MATERIALS AND METHODS
Reagent Amount per reaction
H2O bidest Ad 25µl 10x Taq DNA Polymerase Buffer 2.5µl dNTP mix (dATP, dGTP, dCTP, dTTP) (10mM each) 0.5µl L.CDSB 2‐1 (16.7µM) 0.25µl LCI (16.7µM) 0.25µl Taq DNA Polymerase (5U/µl) 0.25µl DNA‐bead solution 2µl
Description Temperature Time Repetitions Initial Denaturation 95°C 2min Denaturation 95°C 45sec Annealing 60°C 45sec 35 cycles Elongation 72°C 1min Final Elongation 72°C 5min Cooling 4°C
For the second exponential amplification, 5µl PCR product from the first PCR reaction was directly inserted. The linker‐specific and polyU‐specific primers used in this PCR were LCII and E.CDSB 4, respectively.
Reagent Amount per reaction
H2O bidest Ad 50µl 10x Taq DNA Polymerase Buffer 5µl dNTP mix (dATP, dGTP, dCTP, dTTP) (10mM each) 1µl E.CDSB 4 (16.7µM) 0.5µl LCII (16.7µM) 0. 5µl Taq DNA Polymerase (5U/µl) 0.5µl DNA 5µl
Description Temperature Time Repetitions Initial Denaturation 95°C 2min Denaturation 95°C 45sec Annealing 60°C 45sec 35 cycles Elongation 72°C 1min Final Elongation 72°C 5min Cooling 4°C
To analyze the PCR products, 2µl of the amplicons were loaded onto a 2% agarose gel.
47 MATERIALS AND METHODS
2.2.2.12 Biotin Quantification
In order to accurately measure the biotinylation level of labeled DNA in the Tdt‐mediated DSB labeling assay and to determine the optimal amount of biotin‐16‐dUTP, the TermoScientific Fluorescence Biotin Quantitation Kit was used.
The Fluorescence Biotin Quantitation Kit is a colorimetric assay to estimate the biotin‐to‐molecule ratio. It consists of two reagents: HABA (4'‐hydroxyazobenzene‐2‐carboxylic acid) and avidin. In a premix (DyLight Reporter), fluorescent avidin is mixed with HABA, a dye that weakly interacts with avidin. The premix is added to the solution containing the biotinylated DNA. Since biotin has a higher affinity for avidin, biotin displaces the HABA, allowing the avidin to fluorescate. The level of fluorescence is proportional to the amount of biotin in the sample and is measured in a microplate reader by comparing the fluorescence to a biocytin standard curve.
Preparation of a diluted biocytin standard
In order to quantify the biotinylation level in each sample, a standard curve based on a biocytin standard dilution series was prepared and measured in a microplate reader (excitation wavelength: 495nm). First, 1xDPBS (diluent) was prepared by adding 9.5ml ultrapure water to 0.5ml 20xDPBS.
Final biocytin concentration Vial Volume of diluent Volume of biocytin standard [pmol/10µl] A 198µl 2µl of biocytin control 100 B 10µl 40µl of vial A 80 C 20µl 30µl of vial A 60 D 30µl 20µl of vial A 40 E 40µl 10µl of vial A 20 F 45µl 5µl of vial A 10 G 47.5µl 2.5µl of vial A 5 H 50µl 0 0
Preparation of DyLight Reporter Working Reagent (DWR)
To determine the total volume of DWR required the following formula was used:
# # # 90
Immediately before use, 14 parts of 1xDPBS were mixed with 1 part of DyLight Reporter to obtain the calculated amount of DWR solution.
Biotin quantification
For each standard and unknown sample, 10µl triplicates were transferred into a microplate well and 90µl 1xDWR solution was added. After mixing, the plate was wrapped in aluminum foil to protect the solution from light exposure and incubated at RT and 200rpm for 5min in a horizontal shaker. Finally, the microplate was placed in a microplate reader and the fluorescence measured after excitation with a laser at a wavelength of 495nm. Then, a reverse plot of the linear range (10‐60pmol biocytin/10μl reaction) of the standard curve was 48 MATERIALS AND METHODS
prepared and a linear regression equation generated. The resulting equation was used to determine picomoles of biotin in the unknown samples by inserting the sample’s average fluorescence intensity.
2.2.2.13 Linker‐Amplification‐Mediated DSB Trapping (LAM‐DST)
In order to label DSB sites in situ, a ligation approach with DNA linkers of known sequence was developed. First, the DNA break is labeled and subsequently enriched on streptavidin‐coated paramagnetic beads. The captured DNA is then amplified and identified by LAM‐PCR and next‐generation sequencing, respectively.
Fixation and permeabilization
In the first step, 1x106 293T cells transfected with CCR5‐specific ZFN or with the plasmid LV902 as a control for 72h, detached from the 6‐Well plate, washed once with 1xDPBS and centrifuged in a 15ml reaction tube at 1,000rpm and RT for 5min. Then, the cells were fixed with 4% PFA at 37°C for 10min with occasional shaking. Subsequently, the fixed cells were washed with 10ml 1xDPBS before fixation with 0.25% Triton X‐100 solution at RT for 20min. The cells were then washed three times with 1xDPBS and resuspended in 800µl 1xDPBS. The cells were subsequently split into 8 samples (100µl each) into a 96‐Well plate on ice.
Ligation of DNA linker adapters
Before the ligation of the linker adapters to the DSB sites, the 1xDPBS was removed and the cells incubated with 1xT4 DNA ligase or 1xCircLigase buffer (100µl) incubated at 37°C for 15min. During this pre‐incubation step, the ligation mastermix for single‐stranded linker and double‐stranded linker ligation was prepared as follows:
Reagent Amount per reaction 10x CircLigase ligation buffer 3µl
MnCl2 (50mM) 1.5µl ATP (1mM) 1.5µl PEG 8,000 3µl ssGRUV5‐A‐Phthio linker (100mM) 2µl CircLigase ssDNA Ligase (100U/µl) 2µl
H2O bidest 17µl
Reagent Amount per reaction 10x T4 DNA ligation buffer 5µl dsGRUV2 linker (100mM) 1µl T4 DNA Ligase (100U/µl) 2µl
H2O bidest 42µl
The pre‐incubation buffer was removed and the cells resuspended in the ligation mastermix. Ligation of the ssDNA linker was performed at 60°C for 2h, whereas the dsDNA linker ligation was incubated in a PCR cycler at 16°C for 16h. As a ligation control, DNA was replaced with water.
49 MATERIALS AND METHODS
DNA isolation and measuring DNA concentration
For DNA isolation, the Roche High Pure PCR Template Preparation Kit was used according to the manufacturer’s protocol, and the DNA concentration of each sample was determined using the Nanodrop‐1000.
CCR5 locus‐specific PCR
In order to show successful labeling of ZFN on‐target activity sites, two subsequent rounds of nested CCR5 locus‐ specific PCR were performed using the primers CCR5fwd1 and 2 and U5VI bio primers.
Reagent Amount per reaction
H2O bidest ad 25µl 10x Taq DNA Polymerase Buffer 2.5µl dNTP mix (dATP, dGTP, dCTP, dTTP) (10mM each) 0.5µl CCR5 fwd 1 (10µM) 0.25µl U5VIbio (10µM) 0.25µl Taq DNA Polymerase (5U/ul) 0.5µl DNA 200ng
The mastermix was prepared on ice and subjected to the following amplification program:
Description Temperature Time Repetitions Initial Denaturation 95°C 2min Denaturation 95°C 45sec Annealing 58°C 45sec 30 cycles Elongation 72°C 30sec Final Elongation 72°C 2min Cooling 4°C
Following the first 30 cycles, 5µl of the PCR product were subjected to a second round of amplification using nested PCR primers. The mastermix is shown in the table below. The PCR cycling program was identical to the program in the first PCR.
Reagent Amount per reaction
H2O bidest 37µl 10x Taq DNA Polymerase Buffer 5µl dNTP mix (dATP, dGTP, dCTP, dTTP) (10mM each) 1µl CCR5 fwd 1 (10µM) 0.5µl U5VIbio (10µM) 0.5µl Taq DNA Polymerase (5U/µl) 1µl DNA 5 µl
The PCR products were loaded onto a 2% agarose gel for gel electrophoresis. Successful labeling of the CCR5 locus results in the amplification of a 200bp DNA fragment. In addition, to verify site‐specific labeling on the sequence level, the PCR amplicon was cloned into the pCR2.1‐TOPO TA cloning vector and sent to GATC (Konstanz, Germany) for Sanger sequencing.
50 MATERIALS AND METHODS
LAM‐PCR for the detection of genome‐wide DSB
In addition to CCR5 locus‐specific PCR, the DNA samples were subjected to LAM‐PCR. For linear amplification of the linker‐genome junctions, the primers U5V bio and U5VI bio (1.6µM each) were inserted. During nested PCR for exponential amplification, U5VI bio and GRUV5‐2expo (16.6µM each) were used.
2.2.2.14 Pyrophosphate sequencing
Pyrophosphate sequencing of LAM‐PCR products and the SureSelect libraries was performed by GATC (Konstanz, Germany) or the DKFZ Genomic and Proteomics Core Facility (Heidelberg, Germany). For sequencing, the amplicons carrying sequencing‐specific adapters were mixed in specific ratios according to the desired sequencing depth (for 454: 20ng = 1,000 sequences; for Illumina MiSeq: 3µg = 1x107 sequences).
2.2.2.15 Cloning of PCR amplicons using the TOPO‐TA Cloning Kit
For fast and efficient cloning of single PCR products, the commercial TOPO‐TA‐Kit was used. It is based on the observation that DNA polymerases do not create blunt‐ends during PCR, but introduce single adenosine bases at the 3’ end of the PCR product [80]. Hence, the linearized cloning vector, pCR2.1‐TOPO TA, has two single thymidine overhangs at the 3’ ends, enabling complementary pairing of the vector and PCR product. Two molecules of the topoisomerase from vaccinia virus are covalently attached to the vector. The topoisomerase has a ligation activity, which mediates the ligation of the PCR product into the vector. During this reaction, the topoisomerase is released from the vector and a circular vector‐PCR product is formed.
For cloning of PCR products, the following reaction was mixed:
Reagent Amount PCR product 2µl Salt Solution 1µl pCR2.1‐TOPO TA vector 1µl
H2O bidest Ad 5µl
The reagents are mixed on ice and incubated for 5min at RT. Until transformation into chemically‐competent E.coli, the mixture was stored on ice.
2.2.2.16 Transformation of circular DNA into chemically‐competent E.coli
Methods for the introduction of DNA molecules into bacterial cells are called transformation. Here, plasmid DNA molecules were introduced into chemically‐competent E.coli cells by a heat‐shock step. The E.coli strains stbl3 and OneShot TOP10 were purchased from Qiagen.
Prior to transformation, E.coli cells stored at ‐80°C were thawed on ice for 3min. Subsequently, half the volume of cells was transferred into a fresh, sterile 1.5ml reaction tube, kept on ice and also used for transformation. 51 MATERIALS AND METHODS
Next, between 1‐5µl plasmid DNA was added to the cells and incubated for 30min on ice, before being exposed to a 30sec heat‐shock at 42°C. This forces the DNA to enter the cells through either cell pores or the damaged cell wall. The cells are subsequently placed on ice for 2min. Then, 200µl S.O.C. medium was added to each reaction tubes, and the cells incubated at 37°C for 2h, before platting on LB‐agar plates supplemented with ampicillin (100µg/ml). The plates were incubated at 37°C over night
2.2.2.17 Mini‐ and maxipreparation of plasmid DNA
Plasmid preparations are used to extract and purify plasmid DNA from bacteria. According to the size of bacterial culture and plasmid yield, the method is subdivided into mini‐, midi‐ or maxiprep. Here, plasmid preparations from mini and maxi cultures were performed, which involve three steps: growth of bacterial culture, harvesting and lysis and plasmid purification. For mini‐ and maxiprep, the QIAprep spin miniprep and QIAprep spin maxiprep kits were used according to manufacturer’s protocol.
2.2.2.18 Enzymatic DNA restriction digest
Restriction digest is a method to prepare plasmid DNA for analysis and cloning. It takes advantage of naturally‐ occurring enzymes that recognize and cleave DNA at specific sequence sites, thereby generating DNA fragments of specific size or producing compatible ends for cloning. The most commonly‐used restriction enzymes are Type II restriction endonucleases with 4‐8 bp long palindromic DNA recognition sequences. Here, restriction enzymes were used to control cloning efficiency and lentiviral plasmid preparations.
Below, a typical DNA restriction digest is shown:
Reagent Amount Plasmid (1µg/ml) 1µl 10x Digestion buffer 1µl Restriction enzyme 1µl
H2O bidest Ad 10µl
The digestion was performed at 37°C for 2h and the fragments analyzed by gel electrophoresis.
2.2.3 Bioinformatical Methods
2.2.3.1 Automated Sequence Analysis (HISAP)
To map and analyze the sequences obtained from LAM‐PCR and deep sequencing, the raw sequences were processed [81, 82]. To annotate IDLV insertion site informations, the high‐throughput integration site analysis pipeline (HISAP) was used [83]. Starting from raw 454 or Illumina sequences, HISAP removes all vector‐ and LAM‐ PCR‐ specific sequence parts and clusters identical sequences, which are subsequently aligned to the human genome (GRch37/hg19) using BLAT [81] in order to identify the genomic DSB loci. Additional information such as nearby RefSeq genes, their distance and orientation to the integration site, and distance to the closest CpG 52 MATERIALS AND METHODS
island are also calculated. DSB sites from different samples mapping to the same genomic location were called collisions and processed according to their sequence count. If the read count of one the DSB sites was more than four‐times higher than the other, the DSB site with the lower sequence count was deleted. DSB sites with the same sequence count were both excluded from the DSB data set. Further, DSB sites exclusively located in repetitive genomic regions were also excluded.
2.2.3.2 A549 mRNA expression analysis
For A549 gene expression analysis, triplicate microarray data sets were used. These data sets are publicly available at the GEO database server and have the GEO accession numbers GSM661204, 661205 and 661206. Prior to correlation of the expression data with the DSB data set, the mRNA data were bioinformatically processed. First, the expression data were binned according to their expression activity. Therefore, a threshold was set which was twice the background signal intensity (dark corner). Then, the mRNA microarray probes were broadly divided according to their signal intensity into active (>2x threshold) and silent (<2x threshold) transcriptional mRNA signatures. Subsequently, active mRNA signatures were further subdivided into ten bins according to their signal value ranging from low to high expression activity. Moreover, the mRNA expression signatures were correlated to the A549 DSB sites. A DSB site was termed mRNA signature associated when one or more probe signals were found 250kb up‐ or downstream of the DSB site. Further, a DSB site was considered to be located in a transcribed genomic region, if more than four active mRNA signatures were found within 250kb of the DSB site.
2.2.3.3 DNaseI Hypersensitive Sites
DNaseI Hypersensitive Sites (DHS) are genomic regions with increased accessibility for the endonuclease DNaseI, and hence represent open chromatin. For DHS analysis, 117,818 and 372,215 DHS peak data for A549 and NHDF‐A with the GEO sample accession numbers GSM816649, GSM736567 and GSM736520 were obtained from the ENCODE project repository. These data were subsequently correlated with the A549 and NHDF‐A DSB sites. A DSB was considered to be DHS associated when it was located in between the starting and ending position of the DHS.
2.2.3.4 Histone modifications and Transcription Factor Binding Sites
For histone modification analysis, the following A549 and NHDF‐A histone peak data were downloaded from the ENCODE project repository: H3K4me1 (GEO accession numbers: GSM1003495, GSM1003453, GSM1003526), H3K4me2 (GSM1003511, GSM1003496, GSM733753), H3K4me3 (GSM1003542, GSM1003561, GSM733650), H3K9ac (GSM1003544, GSM733709), H3K27ac (GSM1003493, GSM1003578, GSM733662), H3K36me3 (GSM1003494, GSM1003456, GSM733733), H4K20me1 (GSM1003458, GSM1003486), H3K9me3 (GSM1003454, GSM1003553), H3K27me3 (GSM1003577, GSM1003455, GSM733745), H3K79me2 (GSM1003543, GSM1003512, GSM1003554), and H2A.Z (GSM1003580, GSM1003546,GSM1003505). For the CTCF binding site analysis, CTCF binding site peaks for A549 (GSM1003581, GSM1003582) and NHDF‐A (GSM733744) were used. H3K4me1/2/3, H3K9ac, H3K27ac and H4K20me1 are markers for euchromatin, whereas H3K9me3, H3K27me3 53 MATERIALS AND METHODS
and H3K79me2 are considered to be markers for heterochromatin. Genomic regions containing both eu‐ and heterochromatic histone modifications were termed border regions. Histone peaks were correlated with A549 and NHDF‐A DSB sites. A histone modification was considered to be DSB associated when it was found within 2kb of a DSB site. To obtain a DSB enrichment value for every histone modification the average number of histone modifications per DSB site was calculated and the histone frequencies compared to the random DSB data set.
2.2.3.5 Ingenuity Pathway Analysis (IPA)
In order to detect and analyze the enrichment of genes in a specific cohort, network analyses are frequently used. Here, to analyze whether genes in which induced and repaired DSB sites were identified, can be assigned to specific functions or diseases, the program “Ingenuity Pathway Analysis” (IPA) was used. For the analysis, the “NM_Accession” number for each gene marked by IDLV was saved in a separate .txt‐file and uploaded. The resulting networks “Function&Disease” and “Upstream Regulators” with a p‐value smaller than 0.05 were analyzed. In order to calculate the enrichment of specific categories, the frequency of genes associated to the respective category was calculated and divided by the frequency for the random DSB data set.
2.2.3.6 Identification of DSB site clusters in the genome
To investigate, if DSB sites cluster in genomic regions, statistical methods previously described for modeling the non‐random distribution of retroviral integration sites in the genome were applied [84]. A genomic region is termed a DSB hotspot when the number of IDLV‐tagged DSB sites exceeds the number of DSB sites estimated by a uniform distribution of DSB in the genome. Statistically, two, three, four and five captured DSB located in a 30kb, 50kb, 100kb and 200kb window, are considered to be non‐randomly distributed and defined as hotspots. The higher the number of captured DSB is in a maximum of 200 kb genomic window the higher is the intensity and significance of the hotspot.
54 RESULTS
3. RESULTS
3.1 IDLV‐mediated capturing of radiation‐induced DSB sites in vivo
To detect and analyze radiation‐induced DSB sites in vivo, a new method was used to stably mark DSB sites in vivo (Figure 11). Prior to irradiation, target cells are transduced with an integrase‐deficient lentiviral vector, which becomes stably integrated into the genome by the cellular NHEJ‐repair machinery at DSB sites. Thus, the lentiviral vector places a stable genetic tag at the DSB repair site. Cells that successfully repair the induced damage and survive radiotherapy are subsequently expanded in cell culture. In order to analyze the distribution of induced and repaired DSB, the DNA is isolated and subjected to LAM‐PCR and deep sequencing. The resulting vector‐genome amplicons are then bioinformatically processed to obtain the genomic landscape of DSB sites and genomic instabilities (Figure 11).
Transduction DSB induction Genetic taggingof DSB In vitro expansion of survivingcells
FACS
LAM-PCR and Bioinformatical DNA isolation deep sequencing analysis
Non-lethal DNA lesions
Lethal DNA lesions
Figure 11: Experimental scheme for the induction, tagging and detection of radiation‐induced DSB sites in surviving cell populations. Prior to DSB induction by radiation, the cells are transduced with IDLV. Lethal and non‐ lethal DNA lesions become stably marked by IDLV, and the incorporation of IDLV at DSB sites is followed by flow cytometry during expansion of cells. The DSB sites can be amplified and identified by LAM‐PCR and deep sequencing, respectively.
3.1.1 Immunostaining of H2AX foci in irradiated cells
One of the most frequently used methods for the detection and analysis of DSB sites in radiation‐exposed cells is immunostaining of DNA repair proteins using fluorescently‐labeled antibodies. These DNA repair proteins engage and disengage from the DSB site within a 24h period when damaged DNA is sensed and repaired. In order to determine the kinetics and intensity of DSB repair activity following irradiation with 1 and 4Gy photon beams (X‐rays), immunostaining of H2AX foci in A549 and NHDF‐A was performed. In A549 cells, a single 1Gy and 4Gy photon beam led to the formation of 16 and 37 H2AX foci per cell 1h post‐irradiation, respectively. The level of phosphorylation subsequently decreased until reaching background intensity at about 8h after irradiation (Figure 12, Supplementary Figure 1). In NHDF‐A, the same kinetics of phosphorylation and dephosphorylation were observed. However, only 13 and 32 H2AX foci formed 1h after irradiation with 1Gy and 4Gy photon beams (Figure 12, Supplementary Figure 1). Hence, DSB site analysis by immunfluorescence is restricted to the first 8h after irradiation and does not provide any information about cell fate decision. 55 RESULTS
AB 45 45 Control Control 40 40 1Gy Photon 1Gy Photon
35 4Gy Photon 35 4Gy Photon
30 30
25 25
20 20 H2AX Foci per Cell per Foci H2AX Cell per Foci H2AX
15 15
Average Average 10 Average 10
5 5
0 0 024681012024681012 Time post irradiation [h] Time post irradiation [h]
Figure 12: Kinetics of DSB induction and repair by immunostaining of H2AX foci in (A) A549 and (B) NHDF‐A after irradiation with 1 and 4Gy X‐rays. The maximum number of H2AX foci was reached after 1h and terminated at 8h post irradiation. Control: non‐irradiated cells.
3.1.2 MTT assay to determine lethal dose values for etoposide and doxorubicin
As a control for radiation‐induced DSB sites, etoposide‐ and doxorubicin‐induced topoisomerase 2 (TOP2) DSB sites were determined, too. First, TOP2 inhibitor concentrations for efficient DSB induction and cell survival were determined by using the MTT assay. The lethal dose 50 (LD50) values of etoposide and doxorubicin (i.e. the dose at which 50% of the cells die) for Hela and NHDF‐A cells was resolved by incubating the cells with various concentrations of doxorubicin and etoposide, ranging from 5 ‐ 700 nM (doxorubicin) and 0.5 ‐ 150 µM (etoposide) for seven subsequent days. Subsequently, the MTT assay was performed, in which active mitochondria in living cells reduce the tetrazolium dye MTT to the insoluble, purple‐colored formazan. The absorbance at a wavelength of 610nm is proportional to the mitochondrial activity and thus to the number of viable cells (Figure 13, Table 2). 56 RESULTS
A 120 120 y = -0.0863x + 86.373 y = -0.384x + 87.766 R² = 0.8195 R² = 0.8254 100 100
80 80
60 60
40 40 Relative Cell Viability [%] [%] Viability Cell Relative 20 20
0 0 0 200 400 600 800 0 50 100 150 200 Doxorubicin Concentration [nM] Etoposide Concentration [µM] B 140 140 y = -0.9445x + 117.04 y = -21.463x + 87.777 R² = 0.8915 R² = 0.922 120 120
100 100
80 80
60 60
40 40 Relative Cell Viability [%] Viability Cell Relative [%] Viability Cell Relative 20 20
0 0 0 204060800.0 1.0 2.0 3.0 4.0 Doxorubicin Concentration [nM] Etoposide Concentration [µM]
Figure 13: Relative cell viability of NHDF‐A (A) and Hela (B) after seven day treatment with etoposide and doxorubicin (MTT Assay). Absorbance was measured at a wavelength of 610nm in a microplate reader and the cell viability was normalized to untreated control cells. Triplicate measurements for each concentration were performed. From the linear regression, the LD50 values were calculated. Error bars represent the standard deviation of the mean cell viability.
Table 2: Calculated LD50 values for doxorubicin and etoposide in NHDF‐A and Hela.
Cell type LD50 Doxorubicin [nM] LD50 Etoposide [µM] NHDF‐A 421.5 98.4 Hela 70.8 1.8
3.1.3 IDLV‐Delivered DNA‐baits tag radiation‐induced and repaired DSB sites
In order to stably trap radiation‐induced and repaired DSB sites, three cancer cell lines and primary human fibroblasts (NHDF‐A) were transduced with IDLV prior to irradiation with a single dose of photon, proton or carbon beams. Following irradiation, the cells were expanded in cell culture and the frequency of incorporated proviral DNA was measured by EGFP expression from the IDLV vector during passaging. After 25‐30 days which represents six passages in cell culture, stable EGFP expression was observed (Figure 14, Supplementary Figure 2). With the serial passaging of the cells, episomal IDLV vector copies were gradually lost, and the EGFP expression was stabilized by the incorporation of the IDLV vector at radiation‐induced DSB sites in the genome. Irradiation of the human alveolar adenocarcinoma cell line A549 with increasing doses of a photon, proton or carbon beam led to a dose‐dependent increase in the IDLV integration frequency (Figure 14). Moreover, differences in the DSB capture frequency for different radiation sources were observed (Figure 14). In order to investigate whether IDLV trapping is also suitable to capture DSB in different cell types, five different cancer cell lines were transduced with IDLV (MOI 200) and subsequently irradiated with a single beam of carbons, protons or photons. A549 cancer cells showed the highest IDLV integration frequency of all five cell lines with an increase up to 18‐ fold compared to naturally‐occurring DSB after irradiation with 4Gy photon, 4Gy proton or 1Gy carbon beams. In 57 RESULTS
contrast, the A431 epidermal carcinoma cell lines showed the lowest IDLV integration frequency (Figure 14). Hence, IDLV‐mediated DSB trapping enables stable tagging of radiation‐induced and repaired DSB sites in surviving cell populations over time.
Natural 1Gy Photon 4Gy Photon A 79.2% 79 % 79.3%
d0
2% 14.5% 39.7%
d24 FSC
1% 15.7% 38.4%
d34
EGFP 200 B Photon 180 Proton 160 Carbon
140
120
100
80
60
40
20
GFP Enrichment Compared to Non-Irradiated Cells Non-Irradiated to Compared Enrichment GFP 0 0.125 0.25 0.5 1 2 4 10 Radiation Dose [Gy] 28 C Photon 4Gy 24 Proton 4Gy Carbon 1Gy 20
16
12
8
4
GFP Enrichment Compared to Non-Irradiated Cells Non-Irradiated to Compared Enrichment GFP 0 A431 A549 U87 PC3 BxPC3 Cell Type
Figure 14: IDLV‐trapping of radiation‐induced DSB sites in cancer cell lines and human primary fibroblasts. (A) Gradual loss of IDLV‐based DNA baits during in vitro expansion of NHDF‐A. The IDLV integration frequency and loss of episomal vector copies was measured by EGFP expression from viral DNA baits encoding EGFP under the control of the PGK promoter. EGFP expression decreased following irradiation (d0) until reaching a stable percentage in expanded cells (d18 in A549 / d34 in NHDF‐A). The frequency of IDLV incorporation correlates with an increase in DSB induction and NHEJ‐repair. The numbers indicate the frequency of EGFP+ cells. (B) The frequency of DSB trapping is measured by normalizing the level of stable EGFP expression compared to untreated cells. Increasing doses of photon, proton or carbon beams lead to a radiation source‐ and dose‐dependent increase in the IDLV integration frequency up to 180‐ fold higher than naturally‐occurring DSB frequencies in A549. (C) IDLV efficiently labels radiation‐induced DSB sites in various cell types. The difference in DSB trapping efficiencies indicates that NHEJ‐repair activity various among different cell types. FSC: Forward scatterer, EGFP: Enhanced green fluorescent protein
58 RESULTS
3.1.4 IDLV DSB trapping in NHDF‐A with impaired NHEJ‐repair activity
In order to analyze, if IDLV becomes incorporated at radiation‐induced DSB by NHEJ‐repair activity [85, 86], classical and alternative NHEJ‐repair activity were blocked in NHDF‐A by Nu7441 (LD5) and mirin (LD5) prior to irradiation, respectively (Figure 15). During expansion of irradiated cells, the frequency of EGFP+ cells with impaired NHEJ‐repair was significantly increased compared to NHDF‐A with active NHEJ‐repair (Figure 15). A significant increase in EGFP+ cells was observable from passage 3 after irradiation, suggesting that back‐up DNA repair mechanisms (over‐) compensate for the NHEJ‐repair defect in NHDF‐A. In non‐irradiated control cells with impaired NHEJ‐repair activity no increase in the frequency of EGFP+ cells compared to NHEJ‐proficient cells was observed (Supplementary Figure 3), indicating that impairment of NHEJ‐repair alone does not lead to increased DNA damage levels and more IDLV integration events.
A B 40 +Mirin * 35 +NU7441 IDLV Transduction * +Mirin +NU7441 30 Control 25 * *
NHEJ Inhibition [%] Cells 20 * + * * * * * 15 EGFP 10 Irradiation with 4Gy X-Rays 5 Expansionin vitro 0 FACS p3, p4, p6, p8 DNA Isolation LAM-PCR p3 p4 p6 p8
Figure 15: Chemical inhibition of NHEJ‐repair activity increases integration of IDLV in irradiated cells. (A) Experimental scheme: NHDF‐A were transduced with IDLV 48h before irradiation. NHEJ‐repair activity was impaired by Mirin, Nu74441 or both together 3h prior to irradiation. NHDF‐A were expanded for up to eight passages in culture. Integration of IDLV at radiation‐induced and repaired DSB sites was monitored by measuring frequency of EGFP+ cells by flow cytometry. (B) Starting from passage 3, the frequency of irradiated EGFP+ cells with impaired NHEJ‐repair activity was significantly increased compared to irradiated NHDF‐A with active NHEJ‐repair (Control). Error bars represent the standard deviation of the average frequency of EGFP+ cells. Asterix: p‐value < 5x10‐4 (Fisher’s exact test).
3.1.5 Trapping and mapping of TOP2 poison‐induced DSB
As a control for trapping of radiation‐induced DSB, capturing of TOP2 activity sites at which the chemotherapeutic reagents etoposide and doxorubicin block religation of the DNA after DSB induction was performed and compared to radiation‐induced DSB sites. Therefore, Hela cells were transduced with IDLV and subsequently treated with doxorubicin (LD25 and LD50) for seven days. The experiment was cancelled after 14 days, because more than 90% of the Hela cells died or entered senescence following doxorubicin treatment (data not shown). The frequency of EGFP+ cells was not significantly increased compared to untreated Hela cells at day 21 (Supplementary Figure 4). Thus, lower concentrations of doxorubicin and etoposide (LD10, LD5) were used for DSB induction and IDLV trapping. To ensure efficient DSB induction at LD10 and LD5, immunostaining of γH2AX in NHDF‐A was also performed (Supplementary Figure 5). Up to 15 and 16 foci per cell were detected in cells treated with LD5 and LD10 doxorubicin, respectively, compared to four foci per cell in untreated control cells. Hence, doxorubicin induces efficient levels of DNA lesions for IDLV‐trapping. In contrast to doxorubicin, the 59 RESULTS
number of H2AX foci induced by etoposide treatment was not significantly increased compared to untreated control cells. Thus, IDLV‐trapping of TOP2 activity sites was performed using doxorubicin.
NHDF‐A were transduced with IDLV and subsequently treated for seven days with doxorubicin at LD5 and LD10. To distinguish doxorubicin‐induced DSB from naturally‐occurring DSB, DSB sites were also captured in untreated NHDF‐A. The integration of IDLV was monitored by flow cytometry at days 28, 36 and 42 after the beginning of doxorubicin treatment. The percentage of EGFP+ NHDF‐A cells was stable at these three time points (Figure 16). Concentration‐dependent differences in the IDLV‐trapping frequency were also observed. In average, 35.9% and 51.1% of cells treated with LD5 and LD10 doxorubicin were EGFP+, respectively, compared to 1.5% of the untreated control cells. This increase was statistically significant (p‐value < 5x10‐3, Fisher’s exact test), showing that IDLV‐delivered DNA baits are efficient in capturing doxorubicin‐induced DSB sites.
70 Natural LD5 Doxorubicin * 60 LD10 Doxorubicin * * * * * * 50 * * Cells [%] Cells + 40
30
20 Frequency of EGFP
10
0 Day 28 Day 36 Day 42
Figure 16: Frequency of IDLV‐transduced EGFP+ NHDF‐A treated with LD5 and LD10 doxorubicin. Cells treated with LD10 doxorubicin showed the highest IDLV integration frequency. The enrichment of EGFP+ doxorubicin‐treated compared to untreated cells as well as LD5‐treated compared to LD10‐treated cells was statistically significant. Error bars represent the standard deviation of the average frequency of EGFP+ cells. Asterix: p‐value < 5x10‐3 (Fisher’s exact test)
3.1.6 Calculating the number of integrated IDLV copies per irradiated cell
In order to quantify the number of DSB sites marked by IDLV in radiation‐surviving cells, an absolute quantitative real‐time PCR (q‐RT‐PCR) with primers binding specifically to the EGFP locus in the IDLV vector as well as the genomic MYO gene locus were performed (Table 3).
Table 3: Quantification of IDLV integration events. Conc.: Concentration; SD: Standard deviation of the mean Radiation Radiation Mean Conc. Mean Conc. EGFP/MYO Cell type SD MYO SD EGFP Source Dose MYO EGFP ratio Natural 0Gy 68.20 3.75 1.13 0.13 0.017 Photon 4Gy 73.33 4.57 18.93 0.69 0.258 A549 Proton 4Gy 80.7 3.09 20.27 0.42 0.251 Carbon 2Gy 67.7 3.26 14.80 1.69 0.219
60 RESULTS
Since every A549 cancer cell carries two copies of the MYO locus per cell, the number of EGFP copies per cell can be calculated by simply multiplying the EGFP/MYO ratio by 2. In average, 0.034 EGFP copies per cell were identified in non‐irradiated A549 control cells. The number of EGFP copies was increased about 15‐fold to 0.516 and 0.502 copies per cell in photon‐ and proton‐irradiated samples, respectively. A single 2Gy carbon ion beam led to the incorporation of 0.438 EGFP copies per cell, showing that 2Gy carbon irradiation induces similar ionization and DNA damage levels compared to 4Gy photon and proton beams.
3.1.7 SureSelect Target Enrichment for analyzing IDLV vector integrity
Irradiation does not only induce damage to genomic DNA, but also to the lentiviral episomal DNA. Similar to damaged genomic DNA, the broken fragments of the lentiviral vector may be stably integrated at DSB sites in a non‐LTR‐dependent manner. Since q‐RT‐PCR only covers a small portion of the lentiviral vector, the results from q‐RT‐PCR may be an underrepresentation of the total number of marked DSB sites. Moreover, q‐RT‐PCR does not provide any information about the vector integrity after integration. Thus, in order to analyze the IDLV integrity in irradiated cells, the SureSelect Target Enrichment protocol for Illumina Sequencing was applied to genomic DNA isolated from IDLV‐transduced and irradiated NHDF‐A. This method enriches the complete lentiviral vector sequences and enables the identification of IDLV vector‐genome junctions which represent vector breakage sites by pyrosequencing (Table 4).
Table 4: Overview about results of HiSeq sequencing of SureSelect DNA samples. V‐V: vector‐vector; V‐G: vector‐genome Vector Sequence Radiation No. of filtered Correctly vector No. of V‐G No. of reads aligned Count Dose trimmed reads aligned reads junctions reads (V‐G) 0Gy 1,662,774 1,620,487 12,675 10,872 13 203 1Gy 1,990,260 1,981,804 70,102 58,280 99 340 4Gy 1,284,307 1,270,338 64,174 61,228 235 590
About 1% of the filtered and trimmed reads aligned to the vector. Of these sequences, the majority (>95%) were correctly aligned vector‐vector reads, indicating that IDLV is stable upon integration at radiation‐induced DSB. However, 345 vector‐genome junctions in irradiated and non‐irradiated NHDF‐A were also identified (Table 4). In non‐irradiated cells, the majority of vector‐genome junctions were located in the LTR regions and one junction at the packaging signal psi, the RRE as well as the cPPT inside the lentiviral vector (Figure 17). In contrast, vector‐ genome junctions in irradiated NHDF‐A were evenly distributed throughout the vector, showing that the continuity of IDLV was disrupted and that the resulting fragments were integrated into the genome in a non‐LTR‐ mediated manner (Figure 17). 61 RESULTS
A 16 Number ofNumber Baits 0 Baits
IDLV B 25 Irradiated (4Gy) Irradiated (1Gy) 20 Natural
15
10
5 Number of Vector-Genome Number of Vector-Genome Junctions
0 0 500 1,000 1,500 2,000 2,500 3,000 3,500 Position of Vector-Genome Junction in Provirus
Figure 17: Distribution of vector‐genome junctions in IDLV vector (A) Design and distribution of IDLV‐vector baits.No baits were designed for the human PGK promoter sequence in the capture library. (B) Distribution of vector‐genome junctions in NHDF‐A. Vector‐genome junctions in non‐irradiated control cells were identified mainly at the LTR. In contrast, vector‐ genome junctions in irradiated NHDF‐A occurred in almost every site inside the vector, indicating that the vector continuity is disrupted upon irradiation. Natural: Non‐irradiated cells However, a more detailed analysis of IDLV integrity and identification of additional DSB sites from this data set could not be performed at this time, since the total number of vector‐genome junctions in irradiated and control cells was too low. The experimental and bioinformatical analysis pipeline for the SureSelect platform is currently under development in our group. Thus, more vector‐genome junctions in both non‐irradiated and irradiated samples may be identified in the future.
3.2 Analyzing early DSB repair events and kinetics of DSB induction at single nucleotide resolution
As shown in the previous chapters, IDLV‐based trapping of radiation‐induced and repaired DSB sites is suitable to identify and analyze the genomic landscape of DSB and genomic instabilities in radiation‐surviving cells. Thus, IDLV‐mediated DSB trapping might also be useful to identify radiation‐induced DSB sites within the first few hours after DSB induction by IR and follow the DSB site distribution over time. In addition, new approaches to label DNA lesions in real time were developed to study DNA repair kinetics. The development of these methods is described in chapters 3.2.3 and 3.2.4. In order to study DSB induction and repair kinetics in different cell cycle phases, protocols to synchronize cells in G1, S and in G2 phase were established and are also described in the following sections. 62 RESULTS
3.2.1 Genomic tagging of early‐repaired radiation‐induced DSB sites by IDLV
To analyze early DSB repair events, i.e. within 24h after irradiation, and to trace DSB sites over time, NHDF‐A were transduced with IDLV 48h prior to irradiation. Following irradiation with 4Gy X‐rays (photon), the DNA from irradiated and non‐irradiated control cells was isolated at 1h, 4h, 8h, and 24h post irradiation. LAM‐PCR and deep sequencing identified a total of 1731 unique DSB sites in irradiated and control cells (Table 5).
Table 5: Number of DSB sites identified at 1h, 4h, 8h and 24h after irradiation
Radiation Source 1h 4h 8h 24h Photon 137 48 71 101 Natural 1183 49 71 71
The number of unique DSB identified in irradiated NHDF‐A was similar to the number of DSB in non‐irradiated control cells at all four time points after irradiation. A possible reason for this observation is that the majority of IDLV copies in the cell were not yet integrated into the genome, but present as episomes. These vector molecules also become amplified during LAM‐PCR and are thus overrepresented in the sequences. Indeed, only 30% of the sequences that carried the correct sequencing megaprimer sequence and LTR sequence parts also mapped to genomic locations. The remaining 70% were not related to genomic sequences, but to lentiviral vector parts. Moreover, only 1% of the sequences mapping to the genome were IDLV integration sites, indicating that the majority of the genomic sequences map to the same locus. Hence, analyzing the DSB site distribution and repair kinetics in NHDF‐A within the first 24h following irradiation was not possible.
3.2.2 Synchronization of NHDF‐A and Hela cells in G1, S and G2 phase
In order to study the influence of cell cycle phases on DSB induction and repair kinetics, cells need to be synchronized efficiently. Therefore, protocols to block cell cycle progression in G1 and S phase as well as at the G2/M checkpoint were established. The efficiency of each synchronization method was analyzed by staining the nuclear DNA with propidium iodide (PI) foolowed by flow cytometry. For synchronizing cells in G1 phase, serum starvation was performed. Therefore, NHDF‐A were incubated with DMEM containing 1% FCS instead of 10% for 72h (Figure 18). Up to 93% of the NHDF‐A were arrested in G1/G0 phase. Synchronizing NHDF‐A in S phase was achieved by using a double thymidine block. Cells were cultivated in DMEM containing 10% FCS and 2mM thymidine for 48h. Subsequently, the block was released by incubating cells in DMEM (10% FCS) for 16h before cell cycle was blocked for a second time with DMEM containing 10% FCS and 2mM thymidine for another 72h (Figure 18). Double thymidine treatment enables the synchronization of about 74% of all cells in S phase. Nocodazole is a chemical substance frequently used for blocking entry into mitosis. Therefore, cells were arrested in G2 phase by adding Nocodazole at a concentration of 100ng/ml to the cell culture medium for 72h (Figure 18). At these conditions, 46% of all cells were arrested in G2/M phase. Since nocodazole is an anti‐ neoplastic agent that interferes with microtubule formation, the frequency of dead cells increased up to 6%. 63 RESULTS
G1/G0: 75.4% S: 10.8% G2/M: 12.5% Control Dead: 0.1%
G1/G0: 92.6% S: 3.8% G2/M: 3.2% Serum Starvation Dead: 0.2%
G1/G0: 1.7% S: 73.5% G2/M: 10.4% Double Thymidine Dead: 1.0%
G1/G0: 9.6% S: 5.6% G2/M: 45.7% Nocodazole Dead: 5.5%
Figure 18: Synchronization of NHDF‐A in G1/G0, S and G2/M phase of the cell cycle. The method used for synchronization and the distribution of the cells in the different cell cycle phase is indicated at the right side. The PI fluorescence values on the x‐axis indicate the DNA content. Control: Non‐synchronized cells.
3.2.3 Tdt‐mediated labeling of radiation‐induced DSB sites
Since IDLV‐labeling of DSB sites does not enable the study of early DSB repair events and repair kinetics. Since there is a lack of methods to map DSB with high specificity and resolution genome‐wide, a new method to identify rapidly‐repaired DSB sites is required. Thus, an in situ DSB labeling approach was developed and established. In principal, irradiated cells can be fixed using 4% paraformaldehyde and permeabilized with 0.25% Triton X‐100. Subsequently, the free DNA ends produced by irradiation are labelled by the activity of the terminal deoxynucleotidyl transferase (Tdt), which catalyses the addition of the nucleotide dUTP to the 3' terminus of DNA molecules. In addition to dUTP, modified dUTP covalently linked to biotin (Biotin‐16‐dUTP) are also included in the labeling reaction to enable capturing of the DSB sites by streptavidin‐coated paramagnetic beads. The captured DNA can subsequently be digested by a restriction enzyme and short DNA linkers of known sequence ligated to the DNA overhangs. The DNA fragments are finally exponentially amplified by PCR (Figure 19). 64 RESULTS
AB C 2.5 Fixation of treated cells
2.0
Permeabilization of nucleus 1.5 5000 1.0 In situ polyU labeling of DSB 2000 0.5 DSB
3’ 5’ OH OH 1650 5’ 3’ Bio/pmol Primer] Biotin-DNA Ratio [pmol 0.0
+ dUTP 30 20 15 10 8 6 4 2 + Tdt + B-16-dUTP Total Primer Amount [pmol/10µl]
UUUUUUUUU-16-B 1000 D 50 B-16-UUUUUUUUU 850
40 650 DNA Isolation 500 30 DNA restriction digest Magnetic capture 400 Linker Ligation 20 300
UUUUUUUUU-16-B LK
Biotin Concentration [pmol/10µl]) Concentration Biotin 10 LK B-16-UUUUUUUUU 200
Amplification 0 0.05mM 0.1mM w/o Tdt w/o TdT w/o dUTP Sequencing 100 Biotin-16-dUTP Controls
Figure 19: Tdt‐mediated polyU labeling of radiation‐induced DSB sites. (A) Workflow of Tdt labeling and amplification. (B) MNaseI digestion of genomic DNA in situ shows that fixation of cells with 4% PFA and permeabilization with 0.25% Triton X‐ 100 enables enzymes to enter the cell nucleus. Numbers at DNA ladder indicate the DNA fragment size. The numbers above every lane represent the units MNaseI inserted and the incubation time. (C) Determining the biotin detection limit using the Biotin Quantificatioin Kit from ThermoScientific. The lowest amount of biotin detectable in 10µl reaction volume is 6pmol. Grey area indicates concentrations below detection limit. (D) Efficient in vitro biotin‐dUTP labeling of genomic DNA by Tdt with two different concentrations of biotin‐16‐dUTP. Tdt: Terminal deoxynucleotidyl transferease; w/o: without; DSB: DNA double strand break; M: 1.5kb plus DNA ladder
First, standard reaction conditions were established and evaluated. Thus, different protocols and solutions were tested for fixation and permeabilization. Non‐treated A549 cancer cells were fixed with 4% paraformaldehyde (PFA) and permeabilized with 0.25% Triton X‐100. Subsequently, different units of MNase were added for 5 and 10 minutes incubation (Figure 19). The endonuclease MNase recognizes and cleaves double stranded DNA located in between the nucleosomes [87], thereby reducing the median DNA size to several hundred nucleotides in length. MNase digestion of genomic DNA in fixed cells led to smaller DNA fragments, showing that the reaction conditions chosen for fixation and permeabilization are efficient to enable enzymes to enter and become active in the cell nucleus (Figure 19).
Next, to evaluate the activity of Tdt, the DNA biotinylation level was measured. In order to determine the detection limit of biotinylated DNA, the biotinylated DNA primer SK‐LTR 5bio was diluted (2pmol ‐15pmol per 10µl reaction). Since every DNA primer molecule carries a single biotin molecule, the ratio of biotin‐to‐DNA is 1. Hence, the biotin concentration determined is equal to the inserted primer concentration. The biotin detection limit of the biotin quantification assay was the lowest concentration at which the biotin‐to‐DNA ratio was still 1. As shown in figure 32C, the biotin detection limit per DNA molecule was determined to be 6pmol in 10µl reaction volume. Morover, the reaction conditions for the Tdt polymerase were optimized. Therefore, two different biotin‐16‐dUTP concentrations (0.05mM and 0.1mM) were used for in vitro Tdt‐mediated labeling of genomic DNA. Compared to non‐labeled DNA, a 7.6‐fold increase in biotinylation levels was observed (Figure 19). The level of biotinylation in the two samples with 0.05mM and 0.1mM biotin‐16‐dUTP was similar, showing that the chosen concentrations are suitable to label free DNA ends. Furthermore, in situ DSB labeling was performed and the level of biotinylation determined. Therefore, 4x106 293T cells were irradiated with 20Gy X‐ rays, fixed and permeabilized. Subsequently, 1x106 cells were incubated with the Tdt mastermix for 1h at 37°C to label the free DNA ends. Following DNA isolation, the biotinylation level of the DNA was determined. No increase 65 RESULTS
in biotinylation levels compared to the control reaction was observed. By reducing the PFA concentration, introducing more stringent washing steps following permeabilization, incubation with Tdt reaction buffer prior ro DSB labelling and prolonged labelling time did not increase the biotinylation to detectable levels (data not shown), showing that the Tdt activity in situ cannot be measured.
Still, it could be posssible that Tdt activity was efficient to label DSB, but that the biotinylation levels were below the detection limit. Thus, in order to demonstrate that induced DSB sites can be labelled in situ by Tdt, site‐ specific DSB sites were induced in the genome of 293T cells and subsequently labelled. Therefore, 4x106 293T cells were transfected with a ZFN targeting the CCR5 locus [85] and incubated for 24h and 72h. Subsequently, cells were fixed, permeabilized, and the DSB sites labeled by Tdt. Following DNA isolation, LAM‐PCR was performed, resulting in the amplification of DNA fragments in CCR5 ZFN‐treated samples as well as untreated control samples (Figure 20). In order to analyze the specificity of the Tdt‐labeling reaction and amplification, the PCR amplicons were purified and subjected to deep sequencing.
24h 72h M +/+ +/- -/+ -/- +/+ +/- -/+ -/- M
600 500 400
300
200
100
Figure 20: Gel picture of LAM‐PCR amplicons. Numbers at the side of DNA ladder and on top of the gel picture indicate the amplicon size and the time of incubation after transfection, respectively. +/+: +CCR5 ZFN / +Tdt labeling; +/‐: +CCR5 ZFN / without Tdt labeling; ‐/+: without CCR5 ZFN / +Tdt labeling; ‐/‐: without CCR5 ZFN / without Tdt labeling; Roche: negative control (human genomic DNA); H2O: water control; M: 100bp DNA ladder
Table 6: Summary of sequencing results from LAM‐PCR of CCR5‐ZFN treated and Tdt‐labeled DSB samples. ZFN: Zinkfinger Nuclease; Seq: sequence; w/o: without; Tdt: Terminal deoxynucleotidyl transferase, #: Number #Megaprimer #DSB On‐ #DSB Off‐ Sample Tdt Labeling #Raw Seq #DSB found target target CCR5 ZFN +Tdt 8531 8475 48 0 0 CCR5 ZFN w/o Tdt 5562 5516 40 0 0 Untreated +Tdt 1299 1283 29 0 0 Untreated w/o Tdt 4608 4590 53 0 0
The CCR5 locus‐specific ZFN is known to bind and induce a DNA double‐strand break at position 46,414,558 on chromosome 3. LAM‐PCR‐mediated Tdt‐labeling site analysis of the CCR5 ZFN‐treated samples showed that no Tdt‐label was identified at the intended on‐target and off‐target loci [85]. All of the sequences identified were 66 RESULTS
unrelated genomic sites, revealing a high level of unspecific Tdt‐labeling and/or DNA amplification during LAM‐ PCR. Sequence analysis of the amplified genomic sequences revealed that more than half of the sequence carried a thymidine nucleotide at the first two positions after the junction. An increase in the frequency of the two nucleotides adenine and thymidine and a depletion of cytidine as compared to the frequency of these nucleotides in the human genome (A=T=0.29; G=C=0.21) was observed (Supplementary Figure 17). These sequencing results suggest that unspecific DNA sequences were amplified during LAM‐PCR. Hence, in situ DSB labeling by Tdt polymerase does not seem to be efficient for the identification of induced DSB sites.
3.2.4 Linker‐Amplification‐Mediated DSB‐Trapping for DSB labeling in real‐time
Since Tdt‐mediated DSB labeling did not efficiently identify induced DSB sites in situ and suffered from high amplification rates of unspecific DNA sequences, a new method based on direct ligation of DNA adaptors in situ was developed and termed Linker‐Amplification‐Mediated DSB‐Trapping (LAM‐DST). To implement LAM‐DST, 293T HEK cells were transfected with the Zinkfinger nuclease targeting the CCR5 locus [85]. Following a 48h incubation, the cells were fixed with paraformaldehyde (PFA) to stabilize chromatin and prevent artificial DSB formation and subsequently permeablized with Triton X‐100 on ice. Next, the fixed and permeabilzed cells were pre‐incubated with 1x CircLigase ssDNA Ligase reaction buffer or 1xT4 DNA Ligase buffer for 15min to allow the buffer to enter the nucleus. The ZFN‐induced DSBs were directly ligated to DNA linkers that are covalently linked to biotin at 60°C for 2h (CircLigase) or 12°C for 16h (T4 DNA ligase) (Figure 21). The DNA linkers were designed to be compatible with (nr)LAM‐PCR (Supplementary Figure 19). Additional modifications were included in the DNA linker. The 5’ terminus was phosphorylated to enable linker ligation at free 3’OH DNA termini. The 3’ DNA terminus of the DNA linker carries a dideoxycytidine, a pyrimidine analogue, which prevents linker‐linker ligations at the 3’ DNA termini. Furthermore, the phosphate backbone of the last three nucleotides at the 3’ DNA terminus were replaced with a thiophosphoester. Following DNA linker ligation, the genomic DNA was isolated. To show that LAM‐DST labels CCR5 ZFN‐induced DSB, a CCR5 locus‐specific PCR with linker‐ and locus‐specific PCR primers was performed. A PCR fragment of expected size (210bp) was amplified (Figure 21). Cloning of the 210bp PCR fragment into the pCR2.1‐TOPO TA vector and Sanger sequencing verified the amplification of the linker‐genome junction at the CCR5 locus. 67 RESULTS
AB +ZFN -ZFN Fixation of treated cells
Permeabilization of nucleus CCR5 fwd Rep1 CCR5 fwd Rep1 CCR5 fwdRep2 M Rep2 fwd CCR5 In situ linker ligation DSB
3’ 5’ OH OH 5’ 3’
OH P ddC 800 700 600
ddC P OH 500 400 P ddC CCR5 rev primer 300 CCR5 fwd primer ddC P 200
100 Amplification and Sequencing
Figure 21: LAM‐DST for in situ labelling of DSB sites. (A) Workflow and principal of in situ ligation by LAM‐DST. (B) CCR5‐locus specific PCR using the primers indicated in (A) shows that LAM‐DST is suitable to label and identify CCR5 ZFN‐induced DSB sites. Numbers at DNA ladder represent the DNA size. OH: hydroxyl group at 3’ DNA terminus; P: phosphorylated end of DNA linker; ddC: dideoxycarbon at 3’ DNA terminus; M: 100bp DNA ladder; fwd: forward primer; rev: reverse primer; Rep: replicate; ZFN: CCR5‐locus specific Zinkfinger nuclease
In addition to CCR5 locus‐specific PCR, LAM‐PCR and deep sequencing were performed on ssDNA labelled samples to show that LAM‐DST enables the genome‐wide identification of induced DSB sites. The LAM‐PCR was performed with the restriction enzymes MluCI and MseI and showed a polyclonal DNA band pattern (Figure 22). Subsequently, the DNA was subjected to pyrosequencing. 68 RESULTS
+ ssDNA Ligase w/o ssDNA Ligase Controls Expo-PCR Expo-PCR st nd O linear PCR O 1 O 2 2 2 2 H CCR5 ZFN (TransIT) CCR5 ZFN (TransIT) CCR5 ZFN (PEI) CCR5 ZFN (PEI) LV902 (TransIT) LV902 (TransIT) LV902 (PEI) LV902 (PEI) CCR5 ZFN (TransIT) CCR5 ZFN (TransIT) CCR5 ZFN (PEI) CCR5 ZFN (PEI) LV902 (TransIT) LV902 (TransIT) LV902 (PEI) LV902 (PEI) Roche H M MMH
1000 MseI 500 400 300 200
100
1000 MluCI 500 400 300
200 100
Figure 22: LAM‐PCR on ssDNA linker‐labelled DSB sites in CCR5 ZFN‐treated 293T cells. Transfection of CCR5‐specific nuclease was performed with either TransIT or PEI. As a control for the nuclease, the plasmid LV902 that does not encode any nuclease, was also transfected into 293T cells. LAM‐PCR was performed with MseI and MluCI endonucleases. A polyclonal DNA pattern was observed for samples that have been treated with the ssDNA ligase. Unspecific amplification was observed in samples that were not labeled with the ssDNA ligase. ZFN: Zinkfinger nuclease; PEI: Polyethylene‐inositol; M: 100bp DNA ladder, w/o: without; Roche: genomic DNA; H2O: water control
In total, 203,319 raw sequences were obtained for LAM‐PCR samples transfected with CCR5 ZFN and labeled by ssDNA ligase. Of these, 4,369 linker‐genome junctions were identified. In CCR5 ZFN‐treated samples that have not been labeled by ssDNA ligase 10,954 linker‐genome junctions were obtained. LAM‐DST‐mediated analysis of the CCR5 ZFN‐treated samples revealed that no ssDNA ligase site was identified at the intended on‐target and off‐target loci [85]. All sequences were unrelated to the CCR5 ZFN activity, reflecting a high level of unspecific DSB labeling and/or DNA amplification.
3.3 DSB site distribution in radiation‐surviving and expandable cell populations
3.3.1 Identification of radiation‐induced DSB sites by LAM‐PCR
In order to determine the distribution of radiation‐induced and repaired DSB sites in radiation‐surviving cell populations, IDLV‐DSB trapping was chosen. Following expansion of irradiated cells for more than four passages and subsequent genomic DNA isolation, triplicate LAM‐PCR with the two restriction enzymes MluCI and MseI were performed. For each replicate 1µg DNA was inserted into the linear amplification step. The LAM‐PCR amplicons were subsequently prepared for pyrosequencing. For the doxorubicin‐treated cell populations, duplicate LAM‐PCRs with the restriction enzymes MseI and MluCI using 500ng genomic DNA input were performed for samples at day 36 and 42. The LAM‐PCR amplicons were subjected to MiSeq sequencing in the DKFZ core facility. 69 RESULTS
After sequencing, the linker‐ and vector‐parts were removed bioinformatically from the sequences, and the remaining sequence was mapped to the genome using BLAT. From these sequences, collisions with other sequences and non‐mappable DSB sites in repetitive genomic regions were removed. In total, more than 30,000 unique radiation‐induced and repaired DSB sites in A549, U87, PC3 and NHDF‐A were obtained. In doxorubicin‐ treated NHDF‐A, a total of 7739 unique DSB sites and 267 naturally‐occurring DSB in untreated control cells were obtained. In NHDF‐A with impaired NHEJ‐repair activity, a total DSB data set of 22,382 sites from three passages (p4, p6, p8) was acquired (Supplementary Table 1). In order to determine, if the DSB distribution in radiation‐ surviving cells reflects a random distribution, the unique DSB sites were compared to a set of 5,000 randomly‐ distributed, in silico‐generated DSB sites.
3.3.2 DSB are not enriched on chromosomes, in genes and gene‐regulatory regions
First, the chromosomal distribution of radiation‐induced DSB was determined and compared to naturally‐ occurring and doxorubicin‐induced DSB sites. The number of captured DSB per chromosome correlated with the chromosome size, and did not show any difference to the random DSB locations (Figure 23). This was also observed in the irradiated cancer cell lines A549, PC3 and U87, in irradiated NHDF‐A with impaired NHEJ‐repair activity and in doxorubicin‐treated NHDF‐A (Supplementary Figure 6).
12 A Natural Photon 10 Random
8
6 Irradiation 4
Frequency of DSB [%] Sites 2
0 12345678910111213141516171819202122XY Chromosome
12 B Natural LD5 10 LD10 Random 8
Doxorubicin 6
4
Frequency of DSB Sites [%] 2
0 12345678910111213141516171819202122XY Chromosome
Figure 23: Chromosomal distribution of (A) radiation‐ and (B) doxorubicin‐induced DSB in NHDF‐A. The number DSB sites on each chromosome correlates with the chromosome size, and no enrichment of induced and repaired DSB sites on any chromosome was observed. Natural: DSB sites from untreated cells; Random: randomly‐distributed, in silico‐generated DSB sites; LD: lethal dose Next, the distribution of radiation‐induced and repaired DSB sites with respect to gene‐coding regions was analyzed. In average, the frequency of radiation‐induced and repaired DSB was 51.4% in NHDF‐A compared to 70 RESULTS
47.5% and 44.6% for the cancer cell lines and randomly‐distributed DSB, respectively. Moreover, 5.5% and 2.7% of the DSB sites in irradiated primary fibroblasts and cancer cell lines were located in transcriptional control regions located within 10kb upstream of the transcription start site (TSS), compared to 4.7% for the randomly distributed DSB (Figure 24, Supplementary Figure 7, Supplementary Table 2). In doxorubicin‐treated NHDF‐A, the DSB frequency at TSS was comparable to radiation‐induced DSB in NHDF‐A, cancer cell lines and the randomly distributed DSB. However, an increase of induced and repaired doxorubicin‐related DSB sites in RefSeq genes was observed (56.3%) (Figure 24, Supplementary Table 2). Likewise, the frequency of DSB sites in genes and at the TSS in irradiated NHDF‐A with impaired NHEJ‐repair in all three passages was similar to the DSB frequency in cells with active NHEJ‐repair and to the random DSB frequency (Supplementary Figure 7, Supplementary Table 2). Furthermore, the distribution of DSB sites in cancer cell lines with respect to genomic regions with a high content of CpG dinucleotides, termed CpG islands, was compared. These genomic regions are generally associated with gene regulatory elements such as promoters and enhancers. In irradiated cancer cell lines, the DSB frequency at CpG islands [≤1kb] was 3.2% compared to 2.6% for the randomly‐distributed DSB locations (Supplementary Table 2) and decreased with increasing distance from the CpG island, reflecting a random distribution.
In Gene 14.0 NHDF-A Photon 12.0 Natural Random 10.0
8.0
6.0
4.0
DSB Frequency [%] 2.0
0.0 10-5kb 5-0kb 0-10% 10-20% 20-30% 30-40% 40-50% 50-60% 60-70% 70-80% 80-90%90-100%
Upstream In Gene
14.0 NHDF-A Natural 12.0 LD5 LD10 10.0 Random
8.0
6.0
4.0
DSB Frequency [%] 2.0
0.0 10-5kb 5-0kb 0-10% 10-20% 20-30% 30-40% 40-50% 50-60% 60-70% 70-80% 80-90%90-100%
Upstream In Gene
ALD 14.0 ICLV 12.0
10.0
8.0
6.0
4.0 IS Frequency [%] 2.0
0.0 10-5kb 5-0kb 0-10% 10-20% 20-30% 30-40% 40-50% 50-60% 60-70% 70-80% 80-90%90-100% Upstream In Gene Figure 24: DSB frequency in RefSeq genes and at the transcription start site (TSS). Radiation‐ (top) and doxorubicin‐induced (middle) as well as naturally‐occurring DSB sites in NHDF‐A were not enriched at TSS and in RefSeq genes. In contrast to IDLV, integrase‐competent lentiviruses (ICLV) from an ALD clinical study showed preferred integration in gene‐coding regions (bottom). LD5: Lethal Dose 5 Doxorubicin, LD10: Lethal Dose 10 Doxorubicin, Natural: Untreated control cells, Random: Random DSB data set, ICLV: Integrase‐competent lentivirus; ALD: Adrenoleukodystrophy
71 RESULTS
In order to analyze whether the observed IDLV integration profiles are not mediated by residual integrase activity, the DSB distribution pattern was compared to an integrase‐competent lentivirus (ICLV) data set retrieved from two patients from a X‐linked adrenoleukodystrophy (ALD) clinical gene therapy [88]. The ICLV frequency in RefSeq genes was higher than expected by chance and higher than the IDLV frequency in irradiated or doxorubicin‐treated cells (Figure 24, Supplementary Table 2).
Topoisomerase II poisons bind at specific sites at the interface of topoisomerase‐DNA complex. Previous analysis of the DNA sequence at the doxorubicin‐poisoned TOP2 cleavable complex has revealed that doxorubicin has sequence selectivity for inhibiting TOP2 activity, namely an adenine immediately upstream of the cleaved nucleotide [89]. Therefore the genomic sequence adjacent to the vector‐genome junction was analyzed. Indeed, the frequency of the A nucleotide at the first genomic position at doxorubicin‐induced DSB sites was significantly increased (35%) compared to naturally‐occurring (28%) (p‐value < 0.05, Fisher’s exact test) and randomly‐ distributed genomic DSB locations (31%) (p‐value < 3x10‐6, Fisher’s exact test) (Supplementary Figure 8).
3.3.3 IDLV integration at radiation‐induced DSB sites is mediated by NHEJ‐repair
To further analyze the DNA repair mechanism by which IDLV becomes integrated at transiently‐induced DSB sites in the genome, the LTR deletion signatures of the integrated vector were investigated, since deletions are a typical sign of error‐prone NHEJ‐repair. In average, 31% and 53% of IDLV joined into the genome in untreated and irradiated NHDF‐A, respectively, showed deletions of the LTR. In doxorubicin‐treated NHDF‐A, the frequency of LTR deletions was decreased compared to untreated control cells (26% vs. 39%), and deletions varied from a single base up to 40 bases in length. NHDF‐A with impaired NHEJ‐repair activity also showed signs of error‐prone DNA repair activity. In all three passages, the average LTR deletion frequency in non‐irradiated NHDF‐A with impaired NHEJ‐repair was similar to the frequency of LTR deletions in NHDF‐A with active NHEJ‐repair. The frequency of LTR deletions varied between the different NHEJ‐repair inhibitors. For example, in p4, 64% of the LTR sequences in mirin‐treated NHDF‐A carried at least a single deletion, whereas 49% of the LTR in mirin‐ and Nu7441‐treated NHDF‐A were deleted. The frequency of LTR deletions in NHDF‐A with active NHEJ‐repair in the same passage was 55%, indicating that integration of IDLV is mediated by error‐prone DNA repair activity. In contrast to IDLV, integrase‐mediated lentiviral vector insertions showed LTR deletions only in 2% of all integration events (Supplementary Figure 9).
The 3’ terminus of each strand of unprocessed lentiviral DNA ends in the four nucleotide sequence ‘CAGT’. During infection, the integrase processes the 3’ terminus and removes the terminal ‘GT’ dinucleotide. Since IDLV is integrase‐deficient, the ‘GT’ dinucleotide should remain present at the vector‐genome junctions. As expected, the dinucleotide was frequently found at vector‐genome junctions in non‐irradiated NHDF‐A (61%) and irradiated NHDF‐A (30%). Considering a random DSB site distribution, the ‘GT’ dinucleotide frequency would be 5% (Supplementary Table 3). These results and the integration profile of ICLV shown in chapter 3.3.2 indicate that the incorporation of IDLV at naturally‐occurring and induced DSB sites is not mediated by integrase activity, but by the cellular NHEJ‐repair activity.
72 RESULTS
3.4 Analyzing the influence of transcriptional activity, chromatin status and gene classes and networks on DSB site distribution
3.4.1 Trapping of Radiation‐Induced DSB is not influenced by transcriptional activity
The transcriptional activity of the genome has been reported to influence DSB induction and the genomic accessibility of the DNA repair machinery [14]. Thus, in order to evaluate the influence of the transcriptional activity on DSB site distribution and genomic instability, A549 DSB sites were correlated with mRNA microarray probes located within 250kb up‐ and downstream of the DSB sites. Microarray probes with signal intensities above the threshold (2x background intensity) were considered to be active transcriptional mRNA signatures, whereas mRNA signals falling beneath the threshold were classified as silent transcriptional mRNA signatures.
A B Photon Proton Carbon Natural Random Photon Proton Carbon Natural
11.1 10.6 13.4 15.3 16.9 N.E. 4.9 6.9 6.4 6.8 6.1 10
20 63.5 63.5 68.3 69.1 60.8 DSB Frequency [%] 30 Not Expressed 341 122 126 20 801 Expressed 162 54 56 13 287 40 Both 1613 753 524 130 2883 50 Number of DSB Sites in category C 60 Photon Proton Carbon Natural Random 70 mRNA Expression Level 80 36.6 41.1 41.5 41.1 46.1 53.9 90 58.9 58.5 58.9 63.4
DSB Frequency [%] 100
Transcribed 870 428 293 67 1453 Fold depletion/enrichment Non Transcribed 1246 501 413 96 2518 Number of DSB Sites in category 021
Figure 25: Transcriptional activity does not influence DSB site distribution in irradiated A549 cells. (A) More than 80% of DSB sites were associated to mRNA transcriptional signatures of which the majority of DSB sites (65.7%) was associated to both expressed and not‐expressed mRNA signatures. (B) Radiation‐induced and naturally‐occurring DSB sites are not statistically significant enriched at mRNA signatures of specific activity. Numbers indicate the mRNA expression level. (C) In transcribed genomic regions, the DSB frequency was increased up to 1.25‐fold compared to randomly distributed DSB. For statistics, Fisher’s exact test was used. Asterix: p‐value <0.05; N.E. = Not expressed
Using these criteria, more than 80% of the DSB sites identified in irradiated and cultivated A549 were associated to mRNA transcriptional signatures. Of these, in average, 6.2% were associated exclusively to active and 12.5% exclusively to silent transcriptional mRNA signatures. The majority of radiation‐induced and repaired DSB sites (65.7%) were associated to both active and silent mRNA signatures (Figure 25). In order to determine if DSB sites are more frequently associated with highly active rather than weakly active transcriptional mRNA signatures, active mRNA signatures were further subdivided into ten bins according to their signal intensity. Subsequently, the frequency of DSB sites for each mRNA bin was calculated and normalized to the DSB frequency for the randomly distributed DSB locations. An increase in the radiation‐induced DSB frequency at highly active mRNA signatures was observed (mRNA expression activity ≥50%) (Figure 25). However, this increase was not statistically significant. Further, no statistically significant enrichment or depletion for naturally‐occurring DSB sites at active mRNA signatures was observed. Finally, to analyze whether transcribed genomic regions have a higher probability for DSB induction and repair than non‐transcribed regions, the DSB frequency in transcribed genomic regions was calculated and compared to the randomly distributed DSB data set. A genomic region was 73 RESULTS
found to be transcribed when more than four active mRNA signatures were located 250kb up‐ or downstream of the DSB site. Radiation‐induced DSB showed a minor, up to 1.25‐fold enrichment in transcribed genomic regions compared to the randomly‐distributed DSB locations (Figure 25). Taken together, these results show that the transcriptional activity of the genome at the point of irradiation does not influence DSB site distribution.
3.4.2 The location of radiation‐induced DSB sites is composed of histone modifications defining active chromatin
The accessibility of the genome is regulated by modifying different histone variants. These histone modifications influence DNA repair pathway choice and the activity of the DNA repair machinery [2]. The distribution of radiation‐induced and repaired DSB sites and genomic instabilities with respect to histone modifications and genome accessibility in radiation‐ surviving cell populations has not been analyzed to date. Therefore, the DNase I hypersensitive sites (DHS), histone modifications and transcription factor binding sites (TFBS) at radiation‐ induced DSB site in expanded cell populations were analyzed.
DHS are genomic regions sensitive to cleavage by the endonuclease DNaseI. In these genomic regions, the chromatin is not condensed and the DNA exposed. Generally, these regions are associated to transcriptional activity, since this chromatin state is necessary for transcription factor binding [90]. To investigate whether induced and repaired DSB sites are associated to DHS, DHS data available from the ENCODE project repository were correlated with DSB sites from irradiated A549 and NHDF‐A. A DSB site was termed DHS‐associated, if the DSB site was identified within the start and end position of the DHS. The DSB frequency was then normalized to the random DSB frequency, in order to obtain an enrichment value. In irradiated A549 cancer cells, a 2‐fold higher frequency of DSB sites located in DHS (p‐value < 2x10‐4, Fisher’s exact test) was observed compared to randomly‐distributed DSB locations. In irradiated NHDF‐A, no enrichment of DSB sites in DHS was observed (Figure 26, Supplementary Table 4). In contrast, the average frequency of doxorubicin‐induced DSB sites in DHS was 4.3%, representing a 1.7‐fold increase in DSB sites in DHS compared to the random DSB data set (Figure 26, Supplementary Table 4). In irradiated NHDF‐A with impaired NHEJ‐repair, no significant enrichment or depletion of DSB in DHS was observed at any of the analyzed passages (p4, p6, p8) (Figure 26, Supplementary Table 4). Moreover, the signal intensity of the DSB‐associated DHS did not influence DSB site distribution (Supplementary Figure 10).
74 RESULTS
A B p4 p6 p8 0 Photon Ctr
Proton Mirin A549
Carbon +IR Nu7441 Natural
Irradiation MN A Photon 1 Ctr Natural NHDF- Mirin LD5 -IR LD10 Nu7441
NHDF-A Natural MN Doxorubicin Enrichment in DHS compared to Random 3
Figure 26: Enrichment of radiation‐ and doxorubicin‐induced DSB sites in DNaseI HS. (A) Radiation and doxorubicin‐induced as well as naturally‐occurring DSB sites were significantly enriched in DHS. (B) In NHEJ‐impaired NHDF‐A no significant enrichment of radiation‐induced DSB sites in DHS was observed. For statistics, Fisher’s exact test was used. MN: Mirin+Nu7441
Interestingly, radiation‐ induced DSB in NHDF‐A were not randomly distributed inside DHS. DSB sites showed a preferential localization downstream of the center of the DHS with an enrichment up to 11‐fold as expected by chance (Figure 27). Similarly, doxorubicin‐induced DSB sites were also located downstream of the DHS center more frequently.
12 Photon Natural (IR) LD10 10 LD5 Natural (Dox)
8
6
4
2 Enrichment in DSB in Enrichment Frequency Compared Randomto
0 -75 -50 -25 0 25 50 75 Distance from Center of DHS [bp] Upstream Downstream
Figure 27: Distribution of DSB sites inside DHS in NHDF‐A. Radiation‐ and doxorubicin‐induced as well as naturally‐occurring DSB sites were enriched at the beginning and downstream of the DHS center. LD5: Lethal Dose 5 doxorubicin; LD10: Lethal Dose 10 doxorubicin; IR: irradiation; Natural (IR): DSB sites non‐irradiated control cells; Natural (Dox): DSB sites in control cells not treated with doxorubicin In order to analyze whether DSB sites also formed in proximity to DHS, the radius around the DSB sites was expanded up to 10kb. At radii of 2kb, 5kb and 10kb, no differences in the frequency of radiation‐, doxorubicin‐ induced or naturally‐occurring DSB sites compared to the random DSB locations were observed (Supplementary Figure 11), indicating that no accumulation of DSB sites occurs in vicinity to open chromatin.
To further elucidate the influence of genomic accessibility on DSB distribution and genomic instability, the association of induced and repaired DSB sites to histone modifications was analyzed. Therefore, histone 75 RESULTS
modification data were correlated with the A549 and NHDF‐A DSB sites. A DSB site was considered to be histone‐associated, if the histone modification was located within 2kb around the DSB site. H3K4me1/2/3, H3K9ac, H3K27ac and H4K20me1 are markers for euchromatin, and H3K9me3 and H3K27me3 for heterochromatin. Genomic regions containing both eu‐ and heterochromatic histone modifications were termed border regions. Using these criteria, radiation‐ and doxorubicin‐induced DSB sites in NHDF‐A and A549 were found to be frequently enriched in euchromatin and strongly depleted in heterochromatin. Interestingly, both radiation‐ and doxorubicin‐induced DSB sites were also significantly enriched in border regions of eu‐ and heterochromatin (Figure 28, Supplementary Table 5). In agreement with this observation, CTCF transcription factor binding sites, which mark genomic border regions of eu‐ and heterochromatin, were also enriched for induced and repaired DSB sites. The frequency of doxorubicin‐induced DSB sites at CTCF binding sites was increased 1.9‐fold compared to the random DSB distribution (p‐value < 5x10‐3, Fisher’s exact test) (Supplementary Table 6). In NHDF‐A with impaired NHEJ‐repair activity, the frequency of DSB sites associated to border regions and euchromatin increased during cultivation whereas the frequency of DSB sites in heterochromatic regions slightly decreased over time from passage 4 to passage 8 (Supplementary Figure 12). However, the DSB frequency was below the frequency of the randomly‐distributed DSB locations. A significant association of DSB sites in NHEJ‐impaired NHDF‐A to CTCF‐binding sites was only observed in passage 8 of irradiated and Nu7441‐treated cells (p‐value < 2.8x10‐9; Fisher’s exact test) (Supplementary Table 6).
A Irradiation Doxorubicin A549 NHDF-A NHDF-A Photon Proton Carbon Natural Photon Natural LD5 LD10 Natural Euchromatin Heterochromatin Border Region Fold Change in DSB Frequency Compared to Random
01 3 B Irradiation Doxorubicin A549NHDF-A NHDF-A
Photon Proton Carbon Natural Photon Natural LD5 LD10 Natural
H3K4me1
H3K4me2
H3K4me3
H3K9ac
H3K27ac
H3K36me3
H4K20me1
H3K9me3
H3K27me3
Histone EnrichmentCompared to Random
01 3
Figure 28: Distribution of DSB sites with respect to chromatin and histone modifications. (A) DSB sites in irradiated and doxorubicin‐treated cells were significantly enriched in euchromatin and border regions and depleted in heterochromatin. (B) Histone modifications representing transcriptional silencing (H3K9me3, H3K27me3) were depleted at radiation‐ and doxorubicin‐induced DSB sites in A549 and NHDF‐A, whereas histone modifications representing transcriptional activation such as H3K27ac and H3K4me2 were enriched.
76 RESULTS
Next, the histone modifications associated to DSB sites in radiation‐ and doxorubicin‐surviving cell populations were further analyzed (Figure 28, Supplementary Table 7). In irradiated A549, an enrichment of the histone modifications H3K4me1/2/3, H3K9ac and H3K27ac at carbon‐ and photon‐induced DSB sites was observed (Figure 28). Further, genomic regions marked by H3K9me3, which represents heterochromatin, were depleted for DSB sites in irradiated A549. A similar association pattern of histone modifications was also observed in irradiated and doxorubicin‐treated NHDF‐A (Figure 28). However, the enrichment of euchromatic histone modifications was more pronounced in doxorubicin‐treated NHDF‐A than in irradiated cells. In doxorubicin‐ treated NHDF‐A, all euchromatic histone modifications were enriched (Figure 28). In NHDF‐A with impaired NHEJ‐repair activity, an association of euchromatic histone modifications with radiation‐induced DSB sites was observed as well (Supplementary Figure 13). In NHDF‐A treated with Nu7441, the frequency of the histone modifications H3K4me2, H3K9ac, and H4K20me1 increased at radiation‐induced DSB sites during cultivation, whereas the frequency of heterochromatic histone modifications H3K9me3 and H3K27me3 remained nearly constant (Supplementary Figure 13). The comparison of radiation‐induced with naturally‐occurring DSB in A549 and NHDF‐A revealed that radiation‐induced DSB were less frequently depleted in heterochromatin than naturally‐occurring DSB (Figure 28). In contrast, the level of DSB depletion is more pronounced in heterochromatin of doxorubicin‐treated cells. Moreover, the frequency of doxorubicin‐induced DSB was increased in euchromatin compared to naturally‐occurring DSB in untreated cells (Figure 28).
In the genome, histone modifications also contribute to fine‐tuning of expression levels and genomic structures by combining different histone modifications in the same region. In embryonic stem cells for example, specific promoters at genes involved in the supervision of cell fate decisions and differentiation carry both activating and repressing histone modifications [91] and are termed bivalent promoters. Therefore, the frequency of histone combinations at radiation‐induced DSB sites was analyzed and compared to the random and naturally‐occurring DSB locations (Figure 29).
A Irradiation B Doxorubicin
H3K4me1 H3K4me2 H3K4me3 H3K9ac H3K27ac H3K36me3 H4K20me1 H3K9me3 H3K27me3 H3K79me2 H3K4me1 H3K4me2 H3K4me3 H3K9ac H3K27ac H3K36me3 H4K20me1 H3K9me3 H3K27me3 H3K79me2
H3K4me1 1,00 1,03 1,01 0,97 1,00 1,03 1,11 0,90 1,06 H3K4me1 1,10 1,14 1,17 1,13 1,15 1,23 1,04 1,53 1,23
H3K4me2 0,99 1,03 0,94 0,95 0,96 1,07 0,98 0,92 0,98 H3K4me2 1,28 0,94 1,03 0,94 0,99 0,88 0,96 1,10 0,97
H3K4me3 0,99 1,04 1,03 1,02 0,97 0,97 1,07 1,11 0,97 H3K4me3 1,31 1,25 0,96 0,96 0,98 0,87 0,91 1,08 0,93
H3K9ac 0,94 0,97 1,10 1,06 0,98 0,96 1,05 0,96 0,94 H3K9ac 1,27 1,28 1,27 1,00 1,05 0,95 0,92 1,16 1,01
H3K27ac 0,95 1,00 1,11 1,06 1,00 1,00 1,02 0,92 0,95 H3K27ac 1,29 1,28 1,29 1,32 1,16 1,11 1,15 1,25 1,17
H3K36me3 1,07 1,06 1,08 0,99 1,01 0,97 0,95 0,85 0,91 H3K36me3 1,28 1,29 1,31 1,32 1,31 1,05 1,12 1,24 1,17 Natural Natural
H4K20me1 0,99 1,10 1,12 0,98 1,04 0,99 1,02 1,03 1,01 H4K20me1 1,28 1,27 1,32 1,39 1,37 1,25 1,27 1,17 1,11
H3K9me3 1,07 0,98 1,14 1,02 1,05 1,04 1,02 1,10 0,84 H3K9me3 1,24 1,33 1,37 1,39 1,33 1,35 1,31 1,04 1,33
H3K27me3 0,86 1,00 1,22 0,87 0,99 0,99 1,07 1,02 1,03 H3K27me3 1,40 1,30 1,28 1,42 1,49 1,38 1,30 1,22 0,68
H3K79me2 1,11 1,04 1,08 0,96 0,98 0,92 1,13 0,93 0,99 H3K79me2 1,30 1,31 1,31 1,37 1,36 1,29 1,23 1,31 0,71
Photon Doxorubicin
Figure 29: Combinations of histone modifications at DSB sites in irradiated and doxorubicin‐treated NHDF‐A. (A) Specific histone combinations were enriched at DSB sites. In particular, combinations of H3K4me3 at promoters with other histone modifications in irradiated NHDF‐A were more frequently identified at DSB sites than expected by chance. (B) In doxorubicin‐ treated NHDF‐A, combinations of H3K4me2, me3 and H3K9ac were enriched compared to random and naturally‐occurring DSB. Values in boxes represent the enrichment value for each histone combination compared to the random DSB frequency. Values to the right side of the black boxes are histone enrichments values at naturally‐occurring DSB, and values below the black boxes are the values for radiation‐ (A) and doxorubicin‐induced (B) DSB.
77 RESULTS
At radiation‐induced DSB sites in NHDF‐A, the combinations of H3K4me3 with H3K9ac, H3K27ac, H3K36me3 and H4K20me1 were exclusively enriched. H3K4me3 and H3K27me3 were enriched at both radiation‐induced and naturally‐occurring DSB sites and were reported to mark bivalent promoters of developmental genes (Figure 29). In doxorubicin‐treated NHDF‐A, with the exception of the histone combination H3K79me2 and H3K27me3, all combinations were enriched, and the enrichment values were higher than in untreated control cells (Figure 29). In particular, combinations of H3K4me3 with H3K27me3, H3K9ac with H4K20me1, H3K9me3, H3K27me3, and H3K27ac with H4K20me1, H3K9me3 and H3K27me3 were enriched. Similar to radiation‐induced DSB sites, the respective histone modification combinations mark bivalent promoters, active promoters and transcription start sites. In NHEJ‐impaired NHDF‐A, the combinations of histone marks was stable over the three passages. Enrichment of histone combinations involving H3K27me3 and H3K79me2 with H3K9me3 and H4K20me1 was observed (Supplementary Figure 14). This pattern was similar in all irradiated NHDF‐A treated with NHEJ‐ inhibitors and irradiated cells with active NHEJ‐repair (Supplementary Figure 14). In irradiated A549 cancer cells, a strong association of radiation‐induced and naturally‐occurring DSB to histone combinations of H3K27me3 with H3K4me3, H3K9ac, H3K27ac and H3K36me3 was observed. This pattern was consistent for all radiation sources (Supplementary Figure 15).
Taken together, these results show that the frequency of radiation‐induced DSB sites is elevated in accessible genomic regions and in regions regulating gene expression.
3.4.3 Radiation‐induced DSB in radiation‐survivor cells are enriched in genes and networks regulating cell survival
To date, it remains unknown whether cell survival and growth is influenced by the damage and repair of specific gene classes and/or networks. To address this question, the function of genes which showed DSB labeling was analyzed using the Ingenuity Pathway Analysis (IPA) platform. Only genes with intragenic IDLV labeling were included. In total, 1727 and 2557 genes showed intragenic DSB labeling in photon‐irradiated and doxorubicin‐ treated NHDF‐A, respectively. These genes were compared to 1183 and 1566 genes identified in non‐treated control cells and in the randomly distributed DSB locations, respectively. The gene sets from all groups were assigned to 275 and 386 categories in irradiated and doxorubicin‐treated NHDF‐A, respectively. In order to determine the significance of certain functional categories in the list, the enrichment relative to the total number of genes in the random DSB data set in the respective categories was calculated.
Several gene classes were enriched for radiation‐ and doxorubicin‐induced DSB. In particular, genes involved in the regulation of cell cycle, cell death and gene expression were significantly enriched for both radiation‐ and doxorubicin‐induced DSB sites. Moreover, doxorubicin‐induced DSB also showed a preference to be located in genes associated with DNA replication, RNA post‐transcriptional modification, protein degradation and protein synthesis. By analyzing the subcategories of each functional group, differences in the distribution of radiation‐ induced, doxorubicin‐induced, naturally‐occurring and randomly‐distributed DSB sites became apparent. A strong preference for radiation and docxorubicin‐induced DSB sites to be located in genes involved in cell cycle regulation was observed (p‐value < 0.05; Student’s t‐test). In particular, genes associated to cell cycle progression showed an increased frequency of both radiation‐induced and naturally‐occurring DSB sites. Furthermore, genes regulating G1‐, M‐ and interphase were exclusively enriched for radiation‐ and doxorubicin‐ induced DSB. In irradiated and doxorubicin‐treated NHDF‐A, genes with molecular functions involved in the regulation of cell death were significantly enriched (p‐value < 0.05; Student’s t‐test). In particular, genes involved 78 RESULTS
in regulating apoptosis showed a strong enrichment. This enrichment was exclusive for radiation‐ and doxorubicin‐induced DSB, and did not occur in non‐irradiated cells and the random DSB data set (Figure 30). Compared to the random DSB locations, a significant enrichment of radiation‐induced DSB in genes that function in the regulation of gene expression was observed as well (p‐value < 5x10‐3; Student’s t‐test). Radiation‐induced DSB sites were exclusively enriched in genes that activate expression of DNA and RNA (Figure 30).
A cell cycle progression B interphase interphase mitosis arrest in interphase arrest in interphase interphase of tumor cell lines arrest in cell cycle progression G1 phase arrest in interphase of tumor cell lines arrest in interphase of tumor cell lines M phase G2/M phase G1 phase of tumor cell lines ploidy arrest in G2 phase cytokinesis contact growth inhibition arrest in G1 phase of tumor cell lines ploidy of cells G2 phase of tumor cell lines mitosis of tumor cell lines mitosis of cervical cancer cell lines modification of chromatin arrest in mitosis arrest in mitosis arrest in mitosis of tumor cell lines
Cell Cycle Cell arrest in G2 phase of tumor cell lines
arrest in mitosis of cervical cancer cell lines Cycle Cell abnormal cell cycle interphase of cervical cancer cell lines interphase of ovarian cancer cell lines ploidy of liver cells arrest in mitosis of cervical cancer cell lines ploidy of epithelial cells arrest in interphase of ovarian cancer cell lines Photon replicative senescence of cells Doxorubicin length of mitotic spindle arrest in interphase of cervical cancer cell lines entry into S phase of prostate cancer cell lines Natural Natural mitogenesis of tumor cell lines G2 phase of breast cancer cell lines 0 20406080100120140 0 20406080100120140
cell death cell death apoptosis necrosis necrosis apoptosis of tumor cell lines cell death of tumor cell lines cell death of fibroblast cell lines apoptosis of tumor cell lines cell death of fibroblast cell lines apoptosis of fibroblast cell lines apoptosis of fibroblast cell lines apoptosis of connective tissue cells apoptosis of cervical cancer cell lines cell death of muscle cells cell death of kidney cell lines cell death of lung cancer cell lines cell death of embryonic cell lines cell death of lung cancer cell lines cell death of fibroblasts cell death of germ cells cell death of colon cancer cell lines apoptosis of germ cells apoptosis of muscle cells cell death of male germ cells apoptosis of fibroblasts apoptosis of male germ cells necroptosis of cervical cancer cell lines cell death of heart apoptosis of spermatids cell death of heart cells cell survival apoptosis of heart cell viability Cell Death Cell
Cell Death Cell apoptosis of cardiomyocytes cell viability of tumor cell lines anoikis cell death of lymphoblastoid cell lines cell viability of carcinoma cell lines cell viability of fibroblasts cell viability of embryonic cell lines apoptosis of neuroglia necrosis of skeletal muscle cells neurodegeneration of Purkinje cells apoptosis of endometrial cells cell death of brown adipocytes survival of microglia Photon Doxorubicin apoptosis of endometrial stromal cells cell viability of tumor cell lines cellular degradation Natural cell survival Natural cell viability of cervical cancer cell lines cell death of stem cells 0 50 100 150 200 250 300 350 0 50 100 150 200 250 300 350 400
expression of RNA expression of RNA
transcription transcription transcription of RNA
transcription of RNA expression of DNA
transcription of DNA expression of DNA
activation of DNA endogenous promoter transcription of DNA transactivation
activation of DNA endogenous promoter transactivation of RNA
transcription of mRNA transactivation Gene Expression initiation of expression of RNA Gene Expression transactivation of RNA reinitiation of transcription of RNA Photon Doxorubicin
binding of dsRNA Natural initiation of transcription Natural 0 50 100 150 200 250 0 50 100 150 200 250 300
Figure 30: Ingenuity Pathway Analysis (IPA) of radiation‐ and doxorubicin‐induced DSB sites. Genes in irradiated (A) and doxorubicin‐treated (B) NHDF‐A cells with IDLV integration within the gene‐coding region were classified according to physiological function. Enriched genes were frequently involved in processes regulating cell cycle, cell death and gene expression. Within each category, specific disease and function annotations were enriched compared to the random DSB locations. The x‐axis represents the enrichment of DSB sites in the respective gene function compared to the random DSB sites.
These results indicate that the specific gene classes are repaired more frequently upon genotoxic stress, and that irradiation and doxorubicin treatment affect the same gene categories. Therefore, the distribution of induced and repaired DSB sites might also be non‐random with respect to upstream regulators. Thus, genes with intragenic DSB labeling were also analyzed with respect to upstream regulators and miRNA seed sequences. An association of radiation‐ and doxorubicin‐induced and repaired genes to upstream regulators, including transcriptional regulators, growth factors and microRNA seed sequences in NHDF‐A was observed. Within each type of upstream regulator, specific molecules were identified that regulate the expression of genes marked by radiation‐ and doxorubicin‐induced DSB. In the group of transcriptional regulators, DSB sites were located in genes regulated by the stress‐ and apoptosis‐regulators FOS, TP53 and ERG. MicroRNAs regulate the expression of several target mRNAs by binding to specific sequences and inducing their degradation before translation. 79 RESULTS
IDLV‐marked genes in irradiated and doxorubicin‐treated NHDF‐A were regulated by several microRNA including miR182‐5p, miR185‐5p, miR181, and miR145 (Supplementary Table 8, Supplementary Table 9), which have been reported to be involved in the DNA damage response following irradiation [92]. These results indicate that genes involved in maintaining genomic integrity and cell viability and are regulated by specific molecules are preferentially repaired and/or are selected for during expansion.
3.4.4 Identification and analysis of radiation‐induced DSB sites over time
The clonal distribution of modified cells in clinical gene therapy trials is frequently studied by analysing lentiviral insertion sites (IS) [88]. Since every IS is located at a different position in the genome of each cell, IS represent unique marks for every cell and allow to determine the clonality and fate of genetically modified cells [79]. Thus, the identification of radiation‐induced DSB sites in different passages after irradiation may also enable the tracking of DSB sites over time, reflecting the dynamics in the clonality of irradiated cells.
In total, 975 DSB sites were identified in irradiated NHDF‐A in at least two of the three passages (p4, p6, p8). The majority of these DSB sites (726) were identified in both p6 and p8, and 43 DSB were identified in all three passages. The relative frequency of eight of the DSB sites present in all three passages increased from p4 to p8, one remained unchanged and 34 DSB sites decreased. In irradiated NHDF‐A with active NHEJ‐repair, six DSB sites were identified, of which five decreased. In irradiated NHDF‐A with impaired NHEJ‐repair, the majority of DSB sites also decreased over time (Supplementary Table 10). The strongest increase of a single DSB was observed in mirin‐treated and irradiated NHDF‐A. The relative frequency increased more than 21‐fold from p4 (0.019%) to p6 (0.411%) and was located 14kb upstream of the transcription start site of the MED14 gene. This genomic region is characterized by the presence of H3K4me1, H3K4me2, H3K4me3 and H3K27me3 histone modifications. The combination of H3K4me2 and H3K27me3 is associated with poised promoters at embryonic stem cell genes [91]. The strongest decrease in the relative DSB frequency was observed for a naturally‐occurring DSB site in non‐irradiated NHDF‐A with active NHEJ‐repair activity, located more than 615kb downstream of the DIAPH3 gene (0.12% in p4 and 0.001% in p8) and was associated with the heterochromatin histone mark H3K9me3 (Supplementary Table 10). However, no general association of specific histone modifications with an increase or decrease in DSB sites over time was observable.
Comparison of the ten most prominent DSB sites in irradiated NHDF‐A in passage 4, 6 and 8 is presented in figure 31. In irradiated NHDF‐A with active NHEJ‐repair activity, only a single DSB site which was located in intron 7 of FANCD2 gene was present in two subsequent passages with a low retrieval frequency (2.4% and 1.5%). In passage 8 in irradiated NHDF‐A with active NHEJ‐repair, a single DSB site located in the C12orf63 open reading frame with a relative frequency of 23% was present. In NHDF‐A with impaired NHEJ‐repair activity, no dominance of a DSB site was observed in any passage. The total contribution of the top ten DSB sites in each passage did not exceed 30% of the total DSB sites. In mirin+Nu7441‐treated and irradiated as well as Nu7441‐ treated and irradiated NHDF‐A, one and two DSB sites occurred in two passages within the top ten DSB sites, respectively (Figure 31). The two DSB sites in Nu7441‐treated and irradiated NHDF‐A were associated to a gene involved in calcium ion transmembrane transporter activity (ITPR2) and one with unknown function (TMEM90A). In mirin+Nu7441‐treated and irradiated NHDF‐A, the single DSB site was associated to the gene AMOT, whose function is related to angiostatin binding and receptor activity. 80 RESULTS
+ IR Control + IR +Mirin 100% 100%
50% 50% Retrieval Frequency [% of total [% total sequences] of Frequency Retrieval Retrieval Frequency [% of total [% total Frequency of sequences] Retrieval
0% 0% p4 p6 p8 p4 p6 p8
Rank Rank 1. KIAA0564 4.49 IMMP2L 6.59 C12orf63 23.34 1. PLXDC2 1.83 NFAT5 7.69 FANCL 3.26 2. GATA3 3.12 HIST1H2AH 6.30 ADRBK2 2.67 2. DPP10 1.47 SH3BP4 5.46 PXMP3 2.61 3. MMP13 3.05 FAM84A 4.66 C2orf86 1.77 3. PFTK1 1.41 PTPN12 5.39 RARB 2.34 4. YWHAZ 2.12 SLITRK1 4.19 FANCD2 1.48 4. ATAD1 1.28 FNDC3A 4.93 BARD1 1.85 5. AUTS2 1.75 GSDMC 3.84 UBE2D3 1.45 5. CNTNAP5 1.16 PDGFD 4.23 KLHDC6 1.85 6. ADARB2 1.56 PCGF3 3.10 PTPLA 1.34 6. LPHN2 1.10 SLC4A4 3.61 MED14 1.74 7. MBNL1 1.25 FANCD2 2.40 INTU 1.32 7. KCNH7 1.10 TBC1D1 2.61 JAG1 1.59 8. NTRK3 1.06 KCNMB2 2.20 APTX 1.30 8. ARHGAP6 1.10 MDM1 2.33 ZNF33B 1.58 9. SUFU 1.00 MYH13 2.04 C12orf50 1.17 9. EFCAB1 1.04 C15orf41 1.96 STXBP6 1.57 10. BAI3 1.00 ABR 1.50 SPIN4 1.03 10. FAM19A1 0.98 POU2F1 1.43 EPHA4 1.53 All other IS 793 79.61 1184 63.18 1749 63.14 All other IS 754 87.54 1449 60.36 3919 80.08
+ IR +Nu7441 + IR +Mirin +Nu7441 100% 100%
50% 50% Retrieval Frequency [% Frequency of Retrieval total sequences] Retrieval Frequency [% [% Frequency total of sequences] Retrieval
0% 0% p4 p6 p8 p4 p6 p8
Rank Rank 1. ITGA10 1.18 RFNG 4.07 OR1F1 2.09 1. CNTN3 2.07 BCOR 5.74 TMEM90A 7.44 2. CPEB3 1.18 SRrp35 3.78 CXCR4 1.46 2. ZNF33B 1.97 SPTLC3 4.17 ABCA13 2.75 3. KLHL2 1.18 SPOCK1 3.69 ZNF706 1.37 3. CDH12 1.97 ITPR2 2.82 ZP4 1.83 4. KIF2B 1.11 AMOT 3.57 CDKAL1 1.07 4. WNT7A 1.68 SYT17 2.71 GRIA2 1.73 5. TRIM2 1.58 OR2AT4 2.05 RABGAP1L 1.64 5. CMBL 1.11 PLB1 3.36 AMOT 1.04 6. KCNJ3 1.38 FOXR2 2.00 ITPR2 1.54 6. STAG10.96SLC37A33.20BCHE1.02 7. ANO4 1.28 HACE1 1.88 KIAA1377 1.40 7. FAM47A 0.89 SORL1 1.72 P4HA1 1.00 8. SMYD2 0.99 TMEM90A 1.81 FLRT2 1.39 8. CDH8 0.81 ZNF25 1.46 DPYD 0.96 9. GPR123 0.99 OLFM4 1.56 CASK 1.31 9. AP1S3 0.81 MDM4 1.44 CCDC91 0.95 10. POU5F1B 0.79 IMMT 1.48 BMP6 1.26 10. CBLN2 0.74 PXK 1.35 SAFB 0.95 All other IS 606 85.31 1917 73.79 3121 77.72 All other IS 777 90.01 2141 72.36 4233 88.09
Figure 31: Top ten DSB sites in irradiated NHDF‐A with active (Control) and impaired NHEJ‐repair activity in passage 4, 6 and 8. The names of the DSB‐associated RefSeq genes and the relative retrival frequencies are shown in the tables for p4, p6 and p8. IR: irradiated
In order to delineate genomic factors that potentially influence the changes within the top ten DSB sites, the DSB positions relative to the DSB‐associated gene were investigated. In irradiated NHDF‐A with active NHEJ‐repair, four, three and three of the top ten DSB sites in passage 4, 6 and 8, respectively, were located upstream of the TSS. In NHEJ‐impaired and irradiated NHDF‐A, the number of DSB sites located upstream of TSS either remained constant (mirin), decreased (Nu7441) or increased (mirin and Nu7441) during cellular expansion. Similar results were observed for DSB sites located in genes and downstream of the transcription termination site (Supplementary Table 11). Thus, these results indicate that the changes observed in the top ten DSB site in p4, p6 and p8 cannot simply be explained by the location of the DSB sites in gene‐coding regions.
Therefore, the histone modifications in the vicinity of DSB sites might explain the changes observed in the most prominent DSB sites during cultivation. Therefore, the DSB sites from each passage were correlated with histone modification data. Radiation‐induced DSB sites in NHEJ‐proficient cells increasingly accumulated in border regions and disappeared in heterochromatin. In irradiated NHDF‐A with impaired NHEJ‐repair activity, an increase of the top ten DSB sites in border regions was observed as well (Figure 32). Analysis of the histone modifications in vicinity of DSB sites did not show an association of specific histone modifications with DSB sites. 81 RESULTS
A Euchromatin Heterochromatin Border Region p4 p6 p8 p4 p6 p8 p4 p6 p8 Ctr Mirin Nu7441 Irradiation MN
No. of DSB sites
0 5 10
B Histone Modification H3K4me1 H3K4me2 H3K4me3 H3K9ac H3K27ac H3K36me3 H4K20me1 H3K9me3 H3K27me3 H3K79me2 p4
Ctr p6 p8 p4 p6 Mirin p8 p4
Irradiation p6
Nu7441 p8 p4 p6 MN p8 No. of Histones per DSB
N.A. 3 6
Figure 32: Histone modifications associated with the top ten DSB sites in irradiated NHDF‐A in passage 4, 6 and 8. (A) Radiation‐induced DSB sites in NHDF‐A increasingly accumulate in genomic border regions that marked by both eu‐ and heterochromatin histone modifications. (B) The number of histone modifications at radiation‐induced DSB sites in indicated passages. N.A.: No association of DSB site with histone modification
3.5 Identification of frequently damaged and repaired genomic regions
3.5.1 DSB Trapping in irradiated and passaged cells reveals common regions of radiation‐induced and repaired damage
To date, it remains unknown whether genomic regions with increased vulnerability exist and whether the accumulation of DNA damage in specific genomic regions influences cell survival. Such genomic regions would be expected to show a non‐random accumulation of DNA damage and genomic instabilities. Since immunostaining does not allow stable genetic tagging of induced and repaired DNA damage in surviving cell populations over time, IDLV‐based DSB trapping might identify genomic areas in which DNA damage builds up. To analyze whether clustering of induced and repaired DSB sites occurs, a statistical method was applied which was previously used to analyze the distribution pattern of retroviral integration sites [84]. This methodology identifies the number of DNA damages marked by IDLV integrations in genomic regions with a specific size (Figure 33). Generally, two, three, four and five IDLV copies in a genomic region of 30, 50, 100 and 200kb were considered to be non‐randomly distributed and defined as DSB hotspots of second, third, fourth and higher order, respectively (Figure 33). Photon‐, proton‐, and carbon‐induced DSB sites in a cancer cell line were combined as radiation‐induced DSB sites. Moreover, genomic regions which show accumulation of DSB sites were termed DSB hotspots. If a DSB hotspot only contained radiation‐ or doxorubicin‐induced DSB sites, it was termed radiation‐or doxorubicin‐related DSB hotspot, respectively. Similarly, a genomic region showing 82 RESULTS
accumulation of naturally‐occurring DSB sites only was termed a naturally‐occurring DSB hotspot. In case, both naturally‐occurring and radiation‐/doxorubicin‐induced DSB sites occurred in the same hotspot, the hotspot was called to be overlapping.
By applying these criteria to the DSB data sets from irradiated or doxorubicin‐treated cells, a non‐random distribution of DSB sites became apparent (Figure 33, Table 7). In all three irradiated cancer cell lines, the majority of DSB sites existed as single DSB. In average, only 30% of the DSB sites were clustered. Of these, 17% were located in hotspots with an order of four or higher. In contrast to the cancer cell lines, the clustering of DSB sites upon irradiation and doxorubicin‐treatment was increased in primary fibroblasts. In irradiated or doxorubicin‐treated NHDF‐A, 55.7% and 34.6% of the DSB sites were located in DSB hotspots, respectively. Furthermore, in irradiated NHDF‐A, 38.4% of all DSB sites were clustered in genomic hotspots with orders of four and higher. In NHDF‐A cells that were treated with doxorubicin, 17.0% of the DSB sites were clustered in DSB hotspots of fourth or higher order. For both irradiated and doxorubicin‐treated cancer cells and fibroblasts, no enrichment of DSB clusters on specific chromosomes was observed (Figure 33). In NHDF‐A with impaired NHEJ‐ repair activity, the fraction of DSB sites in hotspots was lower than in irradiated and doxorubicin‐treated cells, but increased from passage 4 to 8 from 4.0% up to 27.5%. Similarly, the fraction of DSB sites in hotspots of fourth or higher order increased over time from 1.8% in passage 4 to 22.1% in passage 8 (Table 7).
Table 7: Overview of radiation‐ and doxorubicin‐induced DSB in hotspots in expanded cell populations. Absolute numbers are the sum of both induced and naturally‐occurring DSB sites in the respective cell type. +NHEJi: cells treated with NHEJ‐repair inhibitors
#DSB in #DSB in Fraction of DSB Fraction of DSB in Cell Type DSB Induction #DSB Hotspots Hotspots in Hotspots [%] Hotspots (Order ≥4) (Order ≥4) A549 Radiation 4655 1508 32.4 652 14.0 PC3 Radiation 2261 654 28.7 210 9.3 U87 Radiation 2873 871 30.6 304 10.6 Radiation 10846 6046 55.7 4169 38.4 NHDF‐A Doxorubicin 7991 2767 34.6 1360 17.0 Radiation (p4) 6187 248 4.0 113 1.8 NHDF‐A Radiation (p6) 8391 1126 13.4 736 8.8 +NHEJi Radiation (p8) 16804 4619 27.5 3713 22.1
83 RESULTS
AB 100%
80%
30 kb 100 kb 60% CIS 2ndnd order HS 2 order HSCIS 44thth order order
50 kb A549 40% 200 kb Frequency
CIS 3rdrd order CIS ≥5thth order HS 3 order HS ≥ 5 order 20%
0% C 12345678910111213141516171819202122XY No of DSB Hotspots Chromosome 100% Irradiation
80%
A549 248 518 520 60%
40% Frequency NHDF-A
20% Irradiation 0% 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 X Y Chromosome NHDF-A 22 32 676 100%
80%
60%
40% Frequency [%] Frequency NHDF-A Doxorubicin NHDF-A 5 52 1158 20%
Doxorubicin 0% 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 X Y Chromosome Naturally-occurring Therapy-related DSB in Hotspots (Order≥4) DSB in Hotspots (Order<4) Single DSB
Figure 33: Distribution of DSB hotspots in A549 and NHDF‐A. (A) Definition of DSB hotspots. DSB sites are indicated by the blue circles. (B) Chromosomal distribution of DSB hotspots in irradiated A549, irradiated NHDF‐A and doxorubicin‐treated NHDF‐A. In A549 cells about one third (32.4%) of the DSB sites were clustered, and 18.4% of the A549 DSB sites accumulated in hotspots of 2nd and 3rd order. In irradiated NHDF‐A, 55.7% of the DSB sites were clustered in genomic hotspots of DSB induction and repair, whereas 17.3% of the DSB were located in DSB hotspot of 2nd and 3rd order. In doxorubicin‐treated cells, 34.6% of the DSB were located in hotspots of DSB induction and repair, and 17.0% existed in DSB hotspots of 2nd and 3rd order. Enrichment of DSB clusters on specific chromosomes was not observed. (C) The majority of DSB hotspots in A549 (90%) was exclusively identified in irradiated cells, whereas the frequency of radiation‐related DSB hotspots in NHDF‐A was 40%. Also, more than 90% of the DSB hotspots identified in NHDF‐A were specific for doxorubicin‐treated cells. HS: Hotspots, No: Number
The majority of DSB hotspots in cancer cell lines (90%) was exclusive for irradiated cells and not identified in untreated control cells. In NHDF‐A treated with doxorubicin, the frequency was increased up to 94%, resembling doxorubicin‐related DSB hotspots. In contrast to irradiated cancer cell lines and doxorubicin‐treated fibroblasts, the frequency of radiation‐related DSB hotspots in NHDF‐A was decreased (40%) (Figure 33). In NHDF‐A treated with NHEJ inhibitors, the frequency of radiation‐related DSB hotspots increased from 25% in passage 4 to 66% in passage 6 and 56% in passage 8 (Supplementary Figure 16). Interestingly, within the radiation‐related DSB hotspots the percentage of DSB hotspots containing solely DSB sites from NHEJ‐impaired NHDF‐A increased from 22% to 62% during passaging (Supplementary Figure 16).
In cancer cell lines, 4.3% of all identified DSB hotspots were also identified in non‐irradiated cells. This frequency was further increased in NHDF‐A, where close to 50% of the DSB hotspots contain both naturally‐occuring as well as radiation‐induced DSB sites. In doxorubicin‐treated NHDF‐A, the frequency of these overlapping DSB hotspots was 1%. These results indicate that radiation‐ and doxorubicin‐induced DNA damage hotspots overlap with naturally‐instable genomic regions (Figure 33). The frequency of overlapping DSB hotspots in irradiated NHDF‐A with impaired NHEJ‐repair remained nearly unchanged over the three passages at about 40%. 84 RESULTS
The 20 highest‐ranked DSB hotspots in radiation‐surviving NHDF‐A were located inside or in close proximity to known tumor suppressor genes (FAT1, CUX1, NEGR1)[93‐95], genes driving tumor progression (TRIM36, ARRB1, TCL6)[96, 97] and genes that are frequently mutated in different types of cancer (KIAA0574, TCL6, TRIM29, MOSPD1)[96‐98]. Similarly, an association of 20 highest‐ranked DSB hotspots in irradiated cancer cell lines with cancer‐relevant genes was observed. Identified genes were associated to cancer (ATP2B4, SKI, TMED3, TLR4, CDK10, PIAS1) [99‐105], show tumor suppressor activity (PPP2A, SDCCAG1) [103, 106] and are involved in the regulation of apoptosis (BRE) [107].
In NHDF‐A, a total of 504 and 1261 DSB sites was associated to radiation‐ and doxorubicin‐related DSB clusters with an order of four or higher, respectively. Of these, 213 (42.3%) and 1093 (86.7%) DSB were intragenic, with the majority being located in introns. In contrast to doxorubicin‐specific DSB clusters in NHDF‐A, a preference for radiation‐related DSB hotspots to be located within 10kb upstream of the TSS and at the 3` end of genes (Figure 34). With respect to the affected gene transcripts, the orientation of the IDLV vector located in radio‐ and doxorubicin‐related hotspots did not show any preference to be located in sense or antisense orientation (Figure 34).
A ABTSS
114 60
12 [%] 50
10 40
8 Orientation
30 Gene
6 Irradiation to
DSB Frequency [%] 20 4 relative
2 10 IDLV
0 +10kb12345678910-10kb 0 Intragenic Segment Sense Antisense 5 60 [%] 50 4
40 3 Orientation
30 Gene
2 to
Doxorubicin
DSB Frequency [%] Frequency DSB 20 relative 1 10 IDLV
0 0 +10kb12345678910-10kb Sense Antisense Intragenic Segment
Figure 34: The genomic distribution of radio‐ and doxorubicin‐related DSB hotspots. (A) Radiation‐related DSB hotspots (top) preferentially occurred within 10kb up‐ and downstream of the gene coding region. (B). With respect to the gene orientation, IDLV in radiation‐ (top) and doxorubicin‐related (bottom) DSB hotspots did not show any preference to be located in sense or antisense orientation of the gene.
3.5.2 Radiation‐related DSB hotspots overlap with genes involved in maintaining genome stability and DNA repair
In irradiated NHDF‐A, 54 radiation‐related hotspots with more than four independent DSB per cluster (order >3) were identified, consisting of 88 genes and 412 DSB. More than 50% of these radiation‐related DSB hotspots 85 RESULTS
were intragenic (34 of 54) and predominantely located in introns. Moreover, IPA analysis of the radiation‐related DSB hotspots revealed that the majority of the 88 genes in these hotspots are linked to epithelial neoplasias (76%) and cancer (88%) (Supplementary Table 12). The top five radiation‐related DSB hotspots in NHDF‐A were located inside or in proximity to the genes ERI1, ELF4, TOP1, ZNF33B and ZYX. Interestingly, with the exception of ZNF33B, genetic screens showed that these genes are involved in the maintenance of genome stability, radiation‐sensitivity and DNA repair. In an RNAi‐based screen, the TOP1 gene that encodes the enzyme topoisomerase 1 has been identified to sensitize cells to irradiation [108]. Further, ERI1 is a 3’ exonuclease that degrades histone mRNA at the end of DNA replication, which is important for maintaining genome stability. Hence, ERI1‐deficient cells show increased histone levels that cause chromosome loss and genomic instability [109, 110]. Another RNAi‐based screen revealed that ELF4 and ZYX are involved in the regulation of homologous recombination repair [111].
In NHDF‐A treated with doxorubicin, 171 doxorubicin‐related DSB hotspots (order >4) were identified, consisting of 389 genes and 1053 DSB. Similar to the radiation‐related DSB hotspots in NHDF‐A, the majority (168 of 171) of the doxorubicin‐related DSB hotspots were intragenic and also predominantly located in introns. The majority of DSB hotspots were located in genes associated to neoplasias (68.4%) and cancer (80.7%) (Supplementary Table 13). The top five doxorubicin‐related DSB hotspots were located inside or in proximity to the genes MBTD1/UTP18, BIRC6, SNX1/PPIB/CSNK1G1, RAC1/DAGLB/KDELR2 and XRCC5. These genes have been reported to play an important role in the regulation of DNA repair and genomic stability. UTP18 is an important regulator of radiation sensitivity and DNA repair. Upon depletion by RNAi, the frequency of HR is decreased [111] and cells show increased ionizing radiation sensitivity [108]. Moreover, BIRC6, frequently overexpressed in several human cancers, interacts with p53, thereby facilitating its degradation and frequently leading to carcinogenesis [112]. The fourth largest doxorubicin‐related DSB hotspot was located in intron 1 of the RAC1 gene. RAC1 belongs to the Ras kinase superfamily, and RNAi‐mediated downregulation of RAC1 leads to reduced cell viability [113]. Furthermore, XRCC5 is the 80kDa subunit of the Ku‐heterodimer that recognizes and binds to DSB sites. In complex with Ku70, XRCC5 is required for efficient interaction of Ku heterodimer with DNA‐PK and the initiation of NHEJ‐repair activity [114]. Interestingly, no doxorubicin‐related DSB hotspot overlapped with radiation‐related DSB hotspots, indicating that the distribution of DSB hotspots is dependent on the DNA damaging agent.
Impairement of NHEJ‐repair activity in NHDF‐A prior to irradiation did not influence the distribution of radiation‐ related DSB sites (Supplementary Table 14). In total, 123 DSB hotspots (order>4) containing 685 DSB sites and 219 genes were identified. The top five radiation‐related DSB hotspots in passage 6 were located in proximity or inside LOC729355, RFNG/ CCDC57, SHCBP1, MID1IP1/ BCOR and SORL1. Even though the function of SHCBP1 is still unknown, it may play a role in signaling pathways governing cellular proliferation, cell growth and differentiation (function by UniProtKB/Swiss‐Prot). BCOR is a transcriptional repressor that inhibits gene expression when recruited to promoter regions and may influence apoptosis [115]. The top five radiation‐related DSB hotspots in passage 8 were located in proximity or inside the genes PPIAL4G, HERPUD2, ZNF804A, EFCAB3/METTL2A and ST6GALNAC5/PIGK. Surprisingly, with the exception of the PIGK gene, the function of these genes are unknown, and none of these genes has been reported to be involved in maintaining genome stability or regulating cell death. Moreover, only a minority of the radiation‐related DSB hotspots in NHEJ‐ impaired cells in passage 6 and 8 were associated to cancer and neoplasia (Supplementary Table 14).
Taken together, these results indicate that the DSB hotspot analysis identifies known and currently unknown genomic regions that are important for the maintenance of genome stability and cell survival after irradiation and chemotherapy. 86 RESULTS
3.5.3 DSB hotspots overlap with eu‐ to heterochromatin border regions
By analyzing the DSB hotspot distribution with respect to chromatin topology, it became apparent that the distribution is influenced by the histone environment. In the irradiated or doxorubicin‐treated cells, the majority of naturally‐occuring, overlapping and radiation‐related DSB hotspots were preferentially located in border regions of the genome (Figure 35). In contrast, all three types of DSB hotspots were not frequently detected in heterochromatic regions. Moreover, both radiation‐ and doxorubicin‐related DSB hotspots in NHDF‐A showed association to H3K27ac histones located within 10kb of the DSB hotspots (Figure 35). This histone modification frequently marks active genomic regulatory elements as annotated by the ENCODE project.
A Irradiation Doxorubicin A549 NHDF-A NHDF-A 100 100 100 Natural Natural Natural Overlapping Overlapping Overlapping Radiation-specific 80 80 Radiation-specific 80 Dox-related
60 60 60
40 40 40
20 20 20 DSB Hotspot Frequency [%] Frequency DSB Hotspot DSB Hotspot Frequency [%] Frequency DSB Hotspot [%] Frequency DSB Hotspot
0 0 0 Euchromatin Heterochromatin Border Region Euchromatin Heterochromatin Border Region Euchromatin Heterochromatin Border Region B
MBTD1/ DSB Sites ERI1 DSB Sites UTP18
RefSeq Genes RefSeq Genes H3K27ac H3K27ac
BIRC6 DSB Sites ELF4 DSB Sites
RefSeq Genes RefSeq Genes H3K27ac H3K27ac
Irradiation SNX1/ Doxorubicin DSB Sites DSB Sites TOP1 PPIB/ CSNK1G1 RefSeq Genes RefSeq Genes H3K27ac H3K27ac
RAC1/ ZNF33B DSB Sites DAGLB/ DSB Sites KDELR2 RefSeq Genes RefSeq Genes H3K27ac H3K27ac
Figure 35: Chromatin topology at radiation‐ and doxorubicin‐related DSB hotspots in A549 and NHDF‐A. (A) The histone environment present in radiation‐ and doxorubicin‐related, overlapping and naturally‐occurring DSB hotspots were composed of chromatin modifications favoring euchromatin and border regions which contain both of eu‐ to heterochromatin marks. More than 50% of the radiation‐related, doxorubicin‐related and naturally‐occurring DSB hotspots in A549 and NHDF‐A were located in chromatin regions containing histone modifications present in eu‐ and heterochromatin boundaries. (B) The most intense radiation‐ and doxorubicin‐related DSB hotspots were in direct proximity (<10kb) to gene regulatory elements marked by the presence of H3K27ac. The genes in vicinity to DSB hotspots are depicted. DSB sites in hotspots are displayed in UCSC genome browser.
The frequency of radiation‐induced DSB hotspots in NHDF‐A increased during expansion in cell culture, whereas the frequency of naturally‐occurring hotspots decreased. Further, the proportion of overlapping hotspots in each passage remained constant (Supplementary Figure 15). This change in the relative frequency of the different hotspot types was consistent thoughout the different chromatin regions. The frequency of radiation‐ related DSB hotspots in eu‐ and heterochromatin as well as border regions increased over time, whereas the 87 RESULTS
frequency of overlapping DSB hotspots remained constant (Supplementary Figure 15). Interestingly, in passage 4, 56% of the radiation‐related DSB hotspots were uniquely identified in NHEJ‐impaired NHDF‐A and the relative proportion of these hotspots increased to 73% in passage 8. Finally, the association of DSB hotspots with specific histone modifications was analyzed. This analysis revealed that the frequency of specific histone modifications changed during expansion of cells and that these changes were mostly pronounced at radiation‐related DSB hotspots. The frequency of H3K4me1/2, H3K27ac and H3K27me3 increased more than 1.5‐fold from passage 4 to passage 8. In overlapping hotspots, the histone frequency did not change significantly, whereas in natural DSB hotspots H3K4me3 and H3K27me3 increased more than 1.5‐fold. In radiation‐related DSB hotspots identified solely in NHEJ‐impaired NHDF‐A, the frequency of all histone modifications with the exception of H3K27me3 increased more 1.7‐fold from passage 4 to 8. The strongest histone enrichment was oberved for H3K79me2 (Supplementary Table 16). 88 DISCUSSION
4. DISCUSSION
For determining the distribution of radiotherapy‐induced DSB sites and genomic instabilities in the genome of surviving cancer cell lines and primary human dermal fibroblasts, a total of more than 20,000 unique DSB sites was analyzed and compared to doxorubicin‐induced and naturally‐occurring DSB sites. The results demonstrate that IDLV trapping and LAM‐PCR are suitable methods for the detection of induced and repaired DSB sites and to analyze the genomic factors associated with the DSB sites.
The results presented in this thesis point to a non‐random distribution of radiation‐induced and repaired DSB sites in the genome of radiation‐survivor cells. Radiation‐ as well as doxorubicin‐induced DSB sites were enriched in regions marked by euchromatic histone marks and depleted in heterochromatin. Moreover, radiation‐ and doxorubicin‐induced DSB sites frequently accumulated in small genomic regions which preferentially form at the border of eu‐ to heterochromatin. Genes located in proximity to these hotspots of DSB induction and repair were associated with cancer formation and other genetic disorders, indicating that IDLV‐mediated DSB trapping is a superior method to identify genomic regions which are functionally important for the maintenance of genome stability and possibly for the development of radioresistance. The results and the development of new approaches for direct DSB detection and sequencing will be further explained and discussed.
4.1 Immunostaining of H2AX is not suitable to detect DSB and genomic instabilities in cancer therapy surviving cell populations
To date, the most frequently used method to study DNA repair activity is immunostaining of DNA repair proteins, such as H2AX. As shown in chapter 3.1.1, the microscopically‐visible foci disassemble from the DSB site within 8h after irradiation and chemotherapy, when DSB repair is complete. This is in good agreement with studies on DSB repair kinetics in different cell types [116, 117]. Even though, a few foci may persist as a consequence of complex DNA damages that are difficult to repair [118]. These rare events cannot be efficiently detected, since the number of foci is close to the background level. Moreover, induced DNA damage is not always detectable by immunostaining as observed in the case of etoposide‐induced DNA damage. No increase in H2AX foci in cells treated with etoposide compared to untreated cells was detectable, indicating that TOP2‐ linked DSB remain undetected by the DDR unless TOP2 is removed enzymatically and the free DNA ends are exposed. It has been reported that only a small subset of all DNA breaks produced by etoposide (0.3%) are sensed by the DNA damage response and activate DNA damage repair [119]. Furthermore, H2AX and other repair factors accumulating at DSB sites do not always colocalize acurately with the DSB site, but can form foci that span for megabases around the DSB site, making it difficult to identify the exact position of the DSB site [120]. In addition, immunostaining of DNA repair proteins does not enable tracking of sub‐lethal DSB sites over time in surviving cells and cannot give any insights into a possible link between DSB induction, repair and cell fate decision [121]. Due to these limitations, there is little information available on whether repaired DSB sites in the genome of surviving cells are distributed randomly in the genome, whether certain genomic or cellular factors influence DSB distribution, and how cancer therapy induces therapy resistance, cellular transformation and carcinogenesis. The analysis of induced and repaired DSB sites in therapy surviving cell populations by placing a stable genetic mark at the DSB site could therefore lead to a better understanding of the mechanisms that initiate and drive DSB‐induced therapy resistance and may lead to the development of improved chemo‐ and radiotherapies.
89 DISCUSSION
4.2 NHEJ‐mediated IDLV integration at DSB sites stably marks DNA damage and repair sites in living cells
In this thesis, radiation‐ and doxorubicin‐induced and NHEJ‐repaired DSB were efficiently marked using an integrase‐deficient lentiviral vector (IDLV). The IDLV carries a point mutation in its integrase gene, which blocks its integrase activity, resulting in the generation of increased levels of episomal vector copies in transduced cells. The lentivirus can only become integrated into the DNA in the presence of a DSB by NHEJ‐repair activity [122]. Following irradiation or doxorubicin treatment of different cell types transduced with IDLV, integration of IDLV can be followed by PGK‐driven EGFP expression using flow cytometry. Irradiated and doxorubicin‐treated cells showed significant increase in both H2AX foci numbers as well as in the frequency of EGFP+ cells compared to non‐treated control cells, demonstrating that the different cancer therapy strategies induce high levels of transient DSB, at which IDLV becomes stably incorporated into the genome. By comparing the frequency of EGFP+ cells amongst various irradiated cancer cell lines, it became apparent that IDLV‐mediated trapping at radiation‐induced DSB is not restricted to a specific cell type, and that the different cell types show diverse IDLV integration frequencies, probably reflecting different NHEJ‐repair activities.
IDLV integration was reported to be mediated by NHEJ‐repair activity [122]. In order to show NHEJ‐mediated IDLV integration, the LTR deletion frequency was analyzed, since NHEJ‐repair results in deletion of DNA bases from free DNA ends. This analysis revealed that up to 53% of the integrated vectors carried single bases to large (20bp) deletions. Compared to the LTR deletion frequency after integrase‐mediated DNA integration (2.2%), this demonstrates that IDLV incorporation is mediated by error‐prone Non Homologous End Joining (NHEJ) repair activity. Which of the two sub‐pathways of NHEJ (classical or alternative) is responsible for integration or what percentage each pathway contributes to IDLV integration could not be determined from the LTR deletion frequencies, since both repair pathways frequently result in DNA deletions [19]. Moreover, it is likely that other, homology‐directed DNA repair pathways contribute to the integration of IDLV into the genome, since impairing NHEJ‐repair significantly increased the frequency of irradiated EGFP+ cells and thus IDLV integration events. A possible explanation is that the inhibition of the NHEJ‐repair pathways shifted the ratio between the different repair mechanisms towards repair via homologous recombination. In support of this theory are several studies showing that cells deficient for classical NHEJ‐repair show increased frequencies of alternative NHEJ [123] and homology‐directed repair activities [124]. Moreover, studies conducted in flies where DNA ligase IV, a gene that is required for ligation of free DNA ends in NHEJ [125], was knocked out significantly increased homologous recombination rates [126]. In order to determine which pathway compensates for NHEJ deficiency, the LTR deletion frequency in NHDF‐A with impaired classical or alternative NHEJ‐repair was analyzed. Since the LTR deletion frequency was comparable to the deletion frequency in NHDF‐A with active NHEJ‐repair, it is likely that NHEJ inhibition was incomplete at the chosen NHEJ inhibitor concentrations or that inhibition of NHEJ could have delayed integration until NHEJ‐repair is fully active again.
IDLV‐based DSB trapping is quantitative and enables comparison of the number of DSB sites between different radiation sources. Quantitative‐RT‐PCR analysis revealed that irradiation of A549 cells with a single 4Gy photon beam led to the incorporation of one IDLV copy in every two cells. High LET irradiation with a single 4Gy proton beam induced the same level of DSB sites. The background IDLV integration frequency was calculated to be 0.034 per cell, i.e. about 1 IDLV integration event in 30 cells. These numbers may be an underestimation of the actual number of IDLV‐labeled DSB sites, since only a fraction of the lentiviral vector was amplified and quantified. Moreover, IDLV episomes may also be subject to radiation damage and non‐LTR mediated integration [122]. Indeed, non‐LTR‐mediated integration events have been reported for integrase‐deficient lentiviral vectors [127], and integrated IDLV showed gross reorganizations including deletions, duplications and 90 DISCUSSION
reorganizations of the vector sequence in human colon carcinoma [128]. LAM‐PCR analysis of integrated HIV‐1 IDLV in Hela and fetal mouse embryo SC‐1 cells has revealed both complete and deleted LTR‐end sequences [129]. Thus, in order to analyze the integrity of IDLV and to check for non‐LTR‐mediated IDLV integration, vector target enrichment (SureSelect, Agilent) for HIV‐1‐derived vectors was performed. In this approach, lentivirus‐ specific RNA probes bind to the integration vector, are enriched on streptavidin‐coated beads and prepared for pyrosequencing. The obtained results indicate that the majority of IDLV was intact, and only a few reads showed that the vector continuity was disrupted. However, the number of reads mapping to the lentiviral sequence was low (1%) so that additional optimizations in the protocol to increase the yield are required. Since implementation of the bioinformatical analysis platform in our group is still ongoing, the target enrichment system will be useful to identify both LTR‐ and non LTR‐mediated integration events and possibly detect structural genomic variations at DSB sites.
Taken together, these results demonstrate that IDLV is a versatile tool to capture and quantify cancer therapy‐ induced and naturally‐occurring DSB genome‐wide in various cell types, which enables the characterization of DSB at previously uncharted resolution and detail in living cells.
4.3 Identification of radiation‐induced DSB sites
In order to better understand, how DSB sites are distributed in radiation‐surviving cell populations and how differential DSB repair activity affects cell fate decision, the radiation‐induced DSB sites marked by IDLV were amplified by LAM‐PCR. In total, more than 20,000 unique DSB sites in cancer cell lines and primary fibroblasts that have survived irradiation were obtained. The analysis revealed that the distribution of DSB and NHEJ‐repair activity sites is not biased towards specific chromosomes or chromosomal compartments. Moreover, radiation‐ as well as naturally‐occurring DSB sites were not enriched in gene coding regions, in transcriptional control regions located within 10kb upstream of the transcription start site and in CpG rich regions. Similar results were obtained in doxorubicin‐treated NHDF‐A. Thus, DSB sites are labeled in all genomic regions, can be efficiently amplified and are not enriched in gene‐coding regions.
The obtained results also indicate that IDLV integration into the genome is unlikely to be integrase (IN)‐ mediated, since the IDLV integration pattern does not correlate to the integration pattern of integrase‐ competent lentiviruses (ICLV). ICLV preferentially integrate into gene‐coding regions, at CpG islands and on gene rich chromosomes 16, 17, 19 and 22 [61, 62]. Besides, it was shown that IDLV becomes integrated in pre‐existing DSB as observed in U2OS cells [130]. The ‘CA’ dinucleotide that is located at both LTR ends and that is removed during IN‐mediated integration was frequently detected in the IDLV vector‐genome amplicons. These results demonstrate that lentiviral integration preferences mediated by integrase play no quantitative role in IDLV capture and that IDLV integration events underly mechanism other than integration by ICLV. Hence, the genomic mapping of DSB reflects the repertoire and spatial distribution of acquired genomic damage in irradiated or doxorubicin‐treated cell populations.
4.3.1 Transcriptional activity before irradiation does not influence DSB site distribution
Actively transcribed genomic regions have been described to be prone to DSB induction by exogenous agents [120, 131, 132], since transcription induces an open chromatin state that exposes the genomic DNA. Moreover, actively transcribed genomic regions colocalize with sites of recurrent, early replicating DNA lesions, termed 91 DISCUSSION
early replication fragile sites (ERFS) [133]. During simultaneous DNA replication and transcription, conflicts frequently arise between the DNA replication and transcriptional machinery, which can result in stalling and destabilization of the transcriptional and replication machinery as well as in DSB induction. Such DNA damages arising in ERFS are a potential source for genomic instability and tumorigenesis [133]. Therefore, the mRNA transcriptional activity in the genome of A549 cancer cells around the DSB site was analyzed. More than 200,000 mRNA microarray transcripts were correlated with 4,655 DSB sites in 500kb genomic windows around the DSB site. No enrichment of radiation‐induced or naturally‐occurring DSB sites with expressed or not‐expressed mRNA signatures was observed. DSB sites were not enriched in highly active mRNA transcriptional signatures as well as in transcribed genomic regions that contain more than four active mRNA signatures. These results indicate that the transcriptional activity at the time point of irradiation does not seem to influence DSB site distribution.
However, it cannot be excluded that transcriptional changes that occur in response to irradiation may influence DSB site distribution. In particular, genomic regions encoding genes that are involved in DDR and become actively transcribed upon DSB induction may be more susceptible to DSB induction by reactive oxygen species or delayed genomic instability [44]. In order to further delineate the influence of transcription on DSB distribution and NHEJ‐repair activity, whole transcriptome sequencing and LAM‐PCR during expansion of IDLV‐transduced, radiation‐surviving cells would need to be performed. This analysis would help to unravel the role of both coding and non‐coding RNA on DSB repair activity and induction of delayed genomic instability.
4.3.2 DSB site distribution is non‐random with respect to the genome accessibility
The basic component of the chromatin is the nucleosome, which is composed of histone octamers, surrounded by 147bp genomic DNA and separated by 10‐15bp nucleosome‐free DNA. The accessibility of the DNA and the structure of the genome in the nucleus are regulated by modifications of the histones at their lysine residues. Each histone modification is associated with activation or repression of transcription and with various genomic features such as promoters, transcribed regions, enhancer and insulators [134]. Large‐scale mapping of histone variations such as by the ENCODE project have emerged as powerful tools to characterize the structure of the human genome in different cell types [134]. At promoters, histone modifications contribute to fine‐tuning of gene expression levels. Inactive promoters are characterized by tri‐methylation of lysine 27 at histone 3 (H3K27me3), H3K9me3 and methylated DNA. Prior to transcription, the DNA promoter sequence becomes more accessible for transcription factor binding due to methylation of lysine 4 at histone 3 (H3K4me2, H3K4me3), acetylation and incorporation of histone H2A.Z into the nucleosomes at the transcription start site (TSS). At gene bodies, histones discriminate between active and inactive conformations [134]. Actively transcribed genes are marked by H3K36me3, whereas inactive genes are associated with H3K9me2/3. Moreover, histone marks correlating with active enhancer regions are associated with H3K4me1/2, H3K27ac and H2A.Z, whereas inactive enhancers are enriched for H3K9me2/3.
The DSB repair activity was reported to be unevenly distributed in the genome with regions devoid of repair proteins [14, 131]. Phosphorylated H2AX occurred preferentially in euchromatic and actively transcribed genomic regions after X‐ray irradiation, and in both eu‐ and heterochromatin at times when these regions were undergoing replication. DSB of endogenous origin were reported to be frequently enriched in sub‐telomers, reflecting replication‐ and transcription‐mediated stress during cell division. These results have led to the hypothesis that heterochromatin blocks H2AX phosphorylation due to its high compaction and limited accessibility. However, it was shown that radiation‐induced DSB in heterochromatin can be efficiently detected 92 DISCUSSION
by the DNA repair machinery. Upon recognition, the local chromatin structure is decondensed at the DSB site in order to enable DSB repair to occur [16]. These findings demonstrate that the initially randomly‐induced DSB sites are processed differently in heterochromatin and euchromatin. Since these studies were performed at early time points after irradiation, no information about the distribution of DSB sites with respect to chromatin status in radiation‐surviving cells is available. In particular, the genomic structure and histone modifications at sites of radiation‐induced DSB and genomic instability have not been characterized yet. Thus, in order to analyze the mechanisms of genomic instability in radiation‐surviving cell populations, the DSB sites were correlated with chromatin accessibility and histone modification data available from the ENCODE project.
Open chromatin regions are hypersensitive to DNaseI digestion, because the chromatin has lost its condensed confirmation. These regions are generally associated with transcriptional activity, since an open chromatin state is required for transcription factor binding. Radiation‐induced DSB in A549, but not radiation‐induced DSB in NHDF‐A were significantly enriched in open chromatin. DSB in doxorubicin‐treated cells also showed a significant enrichment in DHS. The enrichment of induced DSB sites was further supported by analyzing the histone modifications at the DSB sites. Radiation‐induced DSB sites in the two cell types A549 and NHDF‐A were significantly enriched in euchromatin, at border regions of eu‐ to heterochromatin and depleted in heterochromatin. The same distribution pattern was also observed for doxorubicin‐induced DSB sites. By further analyzing the histone modifications associated to radiation‐ and doxorubicin‐induced DSB sites, enrichment and depletion of specific histone modifications became apparent. The heterochromatic histone modification H3K9me3 was depleted at radiation‐ and doxorubicin‐induced as well as naturally‐occurring DSB sites. H3K9me3 marks constitutive heterochromatin present at centromers and telomers, which are important functional elements for the maintenance of genome integrity. DSB in heterochromatin were shown to be more detrimental to cell viability than DSB located in euchromatin, since heterochromatin has functions in chromosome segregation, nuclear organization and regulation of gene expression. If DSB accumulate in heterochromatin and cannot be efficiently repaired, genomic instability and loss‐of‐function can occur. Thus, cells that display increased levels of DSB in heterochromatin (HC) are more likely to undergo apoptosis [135], which is reflected in the reduced DSB frequency in HC of radiation‐ and doxorubicin‐surviving cell populations. The DSB frequency in heterochromatic regions marked by H3K9me3 also gives some insight into the spatial nuclear localization of the DSB sites. Since laminar‐associated domains generally interact with genomic regions marked H3K9me3 [136, 137], it is likely that both radiation‐ and doxorubicin‐induced DSB sites are depleted at the nuclear surface and accumulate more frequently close to the center of the nucleus.
In irradiated and doxorubicin‐treated as well as untreated NHDF‐A the histone modifications H3K27ac, H3K36me3 and H4K20me1 were enriched at identified DSB sites. H3K36me3 is associated with actively transcribed genes, and H4K20 becomes methylated at active gene promoters [138]. In addition, H3K4me1 and me2 were enriched exclusively at DSB sites identified in doxorubicin‐treated NHDF‐A, representing active promoter and enhancer regions. Interestingly, some DSB were marked by specific combinations of two histone modifications. In particular, the frequency of H3K4me3 and H3K27me3 was enriched at radiation‐induced as well as naturally‐occurring DSB. Developmental genes that are silenced in embryonic stem cells and become transcribed during differentiation are marked by H3K27me3 and H3K4me3 at their promoters [91]. Their exact role in differentiation and development remains unknown. However, a recent study showed that these bivalent histone marks define gene sets in ovarian cancer that distinguish malignant, tumour‐sustaining and chemo‐ resistant ovarian tumor cells, indicating that upon damage deregulated promoter activity can induce cancer formation [139]. Furthermore, exclusively in irradiated‐ and doxorubicin‐treated cells, the histone modification H3K4me3 was enriched in combination with several other histone marks including H4K20me1, H3K36me3 and 93 DISCUSSION
H3K27ac at DSB sites. These histone combinations are only present at active promoters and transcription start sites in the genome, indicating that cancer‐therapy induced DSB in surviving cell populations preferentially occur or are select for open chromatin associated with gene regulatory regions. Thus, IDLV‐based DSB trapping identifies functional genomic sites in which error‐prone DSB repair in promoters, genes and at TSS may have consequences for gene regulation and expression activity by directly affecting inter‐ and intrachromosomal interactions.
4.3.3 Genes involved in specific cellular processes are enriched for induced DSB
To date, it has not been evaluated, whether cell survival after therapy is dependent on the repair of specific gene classes. Thus, Ingenuity Pathway Analysis (IPA) was performed to investigate, whether specific gene classes that have been damaged and repaired were enriched in survivor cells. Captured DSB in irradiated cells were frequently and significantly enriched in genes that have specific functions related to the control and regulation of gene expression, cell cycle, cell death and survival, which are relevant factors in oncogenesis. Similarly, doxorubicin‐induced DSB sites also preferentially occurred in genes with the same functions. This indicates that error‐prone repair of cancer‐therapy induced DSB could deregulate the expression of these genes. To which extent these genes were inactivated or activated could not be determined. However, mutations in these gene classes could potentially lead to increased DNA damage induction as well as lack of cell cycle control and defects in apoptosis induction, which in turn could cause further destabilization of the genome [140]. Indeed, several human disorders associated with defects in genome maintenance show enhanced cancer susceptibility (Table 1).
Moreover, both radiation‐ and doxorubicin‐damaged and repaired genes share common upstream regulators and target sequences of specific microRNA. The down‐ and up‐regulation of microRNAs in response to radiotherapy has been shown to be associated with the DNA damage response following the exposure to radiation. Particular micro (mi)RNAs have been identified to regulate DSB‐labeled genes in irradiated NHDF‐A. Some of these miRNA were reported to increase cellular survival following irradiation and altered radiosensitivity of cell lines and primary cells [141, 142]. The miRNA‐185 which was exclusively identified as an upstream regulator in irradiated NHDF‐A is a known inhibitor of ATR transcription, and repression of this miRNA conferred radio‐resistance upon X‐ray exposure [143]. Moreover, miRNA‐145 promotes apoptosis and modulates radiation sensitivity of cancer cells. Among its various target mRNA is the p53‐negative regulator MDM2, which is upregulated in epithelial cancers, thereby inhibiting p53‐induced apoptosis [144]. Taken together, these results indicate that cell survival may be dependent on the repair of specific gene classes and genes that are regulated by stress‐response regulating proteins and miRNA.
4.3.4 Clonality of DSB site distribution over time
The identification of radiation‐induced DSB sites in different passages after irradiation enables the tracking of DSB site kinetics. Because IDLV‐tagged DSB sites uniquely mark individual irradiated cells, the dynamics in DSB distribution reflect the changes in the clonality of irradiated cells [79].
Analysis of the ten strongest DSB sites in irradiated NHDF‐A at three different passages revealed that no dominant DSB clone arose in the course of cultivation. Instead, the results point to a more polyclonal cell population in which a few clones rise, disappear and are quickly followed by other clones. In order to find out what factors influence appearance and disappearance of radiation‐induced DSB sites, the genomic location and 94 DISCUSSION
the histone modifications at radiation‐induced DSB sites detected in all three passages were analyzed. Neither the position of the DSB sites around the transcription start site nor a specific histone modification was associated with rise and decline of top ten DSB sites. This was also true for NHDF‐A cells with impaired NHEJ‐repair activity. Thus, it is more likely that these cells carry additional, non‐labeled DSB sites and genomic aberrations that could induce changes in gene regulation and expression [140]. To determine which structural genomic changes occur and how these influence gene expression more deeply, single cell clones need to be isolated and the genome structure as well as transcriptional activity analyzed. In particular, analyzing the frequency of translocations is highly important since these events are threatening genomic integrity which contributes to both haematopoietic and solid tumors. Recent studies indicate that these translocations are not randomly distributed, but appear to be influenced by the nuclear architecture and the transcriptional activity at the translocation site. However, these studies are non‐conclusive for the whole genome, since these data stem from studies conducted in cells at a single locus. Thus, analyzing the mechanisms underlying the occurrence of genomic rearrangements would help to understand their functional consequences.
4.3.5 Radiation‐induced and repaired DSB sites cluster in hotspots in specific genomic regions
Analyzing the distribution of DSB sites in the genome revealed that radiation‐induced and naturally‐occurring DSB sites clustered in hotspots. The most intense genomic hotspot of DSB in irradiated and expanded cell populations harbored up to 131 DSB in a 200kb genomic region and overlapped with genomically unstable regions in non‐irradiated cells. These regions, referred to as overlapping DSB hotspots, were located inside or in close proximity to genes which have been frequently associated with tumor suppressor activity (FAT1, CUX1, NEGR1), have anti‐angiogenic function (OPTC), are frequently mutated in various types of cancers (TRIM36, FCGR1B, TCL6, SLCO2B1) and other genetic disorders (ARRB1, NFASC, TRIM29, MOCS2). Since the identified overlapping hotspots were associated with cancer‐relevant genes, this suggests that these regions are intrinsically unstable and that cancer‐therapy induced genomic instability increases the mutation level in these regions. Damage and error‐prone repair may result in genome defects initiating or promoting radioresistance and carcinogenesis. This becomes apparent by the findings affecting the FAT1 locus. The tumor suppressor gene is located in a genomic region frequently lost in numerous types of human cancer and is mutated at high prevalence, leading to aberrant Wnt activation [93].
The majority of the identified DSB hotspots in irradiated and doxorubicin‐treated NHDF‐A, however, were specific for radiation‐ or doxorubicin‐related DNA damage, respectively, and not detectable in untreated control cells. The top five radiation‐ and doxorubicin‐related DSB hotspots in NHDF‐A were located inside or in proximity to genes shown to be involved in the maintenance of genome stability (ERI1, BIRC6), regulation of radiosensitivity (RAC1, UTP18, TOP1) and DNA repair (ELF4, ZYX, UTP18, XRCC5). Since these therapy‐related hotspots do not contain naturally‐occurring DSB, this indicates that these genomic regions are not naturally unstable, but that induced genomic stress destabilizes the genome in these locations and causes accumulation of DSB. Similar to the overlapping DSB hotspots, the identified genes may potentially be involved in cancer therapy‐induced carcinogenesis and tumor‐resistance mechanisms which influence survival of sub‐lethally treated cells in clinical therapy as exemplified by the ERI1 and RAC1 genes. ERI1 is an exonuclease responsible for the degradation of histone mRNA upon completion of DNA replication. Failure to degrade histone mRNA was shown to result in loss of genomic fragments and whole chromosomes [109, 110]. However, no functional study has reported that mutations or loss of ERI1 can drive cancer initiation or progression. In contrast, the Rho GTPase Rac1 has been implicated in head and neck squamous cell carcinomas (HNSCC) insensitivity to 95 DISCUSSION
radiotherapy, which frequently results in tumor recurrence. In radiation‐resistant cells increased expression, activity and nuclear translocation of the GTPase protein was observed, and chemical inhibition of Rac1 expression and activity resulted in significant improvement of HNSCC sensitivity to ionizing radiation [145]. In addition to this association, strong and weak genomic enhancers annotated by the ENCODE project in direct proximity (<10kb) to the radiation‐ and doxorubicin‐related DSB hotspots were identified. This may indicate that a potential alteration of long‐range inter‐ and intra‐chromosomal enhancer function and of the nuclear architecture can be expected upon error‐prone repair by disrupting the base composition at genomic enhancer regions. Interestingly, radiation‐ and doxorubicin‐related DSB hotspots occurred in different genomic regions, which did not overlap. This indicates that the susceptibility of genomic regions and the pattern of instabilities may depend on the mechanism of DSB induction, i.e. irradiation or interference with DNA replication. Hence, analyzing the DSB hotspot distribution might therefore be used to predict tumor response to specific cancer treatment. Furthermore, it could be speculated that not only the location of DSB hotspots, but also the frequency and type of mutations induced by irradiation is different from doxorubicin. Indeed, it has been shown that cells that have been irradiated or treated with doxorubicin at the equivalent cytotoxic concentration show more translocations after irradiation than doxorubicin‐treated cells. Accordingly, doxorubicin‐treated cells would be expected to show lower translocations frequency [146].
These results clearly indicate that IDLV‐based DSB trapping identifies known and new genomic regions that are destabilized in radiation‐surviving cell populations and may lead to induction of genome destabilization and radiation‐resistance. Hence, in comparison to previous DSB detection technologies, IDLV‐based DSB trapping enables the identification of functional DSB sites that influence genome stability and cell survival. In order to evaluate experimentally whether the identified genomic regions and genes influence genome stability and cell survival, the rapid developments in the field of genome editing can accomplish this task. Endonucleases such as TALENs or CRISPR/Cas9 systems can be designed to target basically every sequence in the genome and induce DSB, which are subsequently repaired by the cellular DNA repair machinery. These site‐specific nucleases can be used to disrupt the integrity of the DNA at the identified DSB hotspots. The effect of these deletions on genome integrity, apoptosis induction, cell survival and cellular growth under radiation‐ or doxorubicin‐induced stress can be analyzed.
Overlapping as well as radiation‐ and doxorubicin‐related DSB hotspots were associated to euchromatic and border regions of eu‐ and heterochromatin. These genomic regions are not located at the periphery of the nucleus, but closer to the center. In genomic border regions, an increased frequency of persisting DNA damage sites was reported [147]. Border regions of eu‐ and heterochromatin are at high risk for non‐allelic homologous recombination during meiosis when numerous programmed DSB are induced for recombination. Furthermore, heterochromatin boundary regions are hotspots for de novo kinetochore formation and contribute to centromere identity [148]. The kinetochore is a large protein complex associated with centromeric DNA that is responsible for mediating the segregation of sister chromatids during mitosis via interactions with the mitotic spindle. Thus, the protection of border regions appears to be a general requirement to prevent formation of dysfunctional centromeres and self‐induced genome rearrangements that are associated with several genomic diseases [149, 150].
96 DISCUSSION
4.4 Methods for in situ labeling of induced DSB sites
Since IDLV‐based DSB labeling is feasible to study the distribution of induced and repaired DSB sites in cancer‐ therapy surviving cell populations, it is reasonable to assume that it might also enable the analysis of DNA repair kinetics within the first few hours after irradiation. LAM‐PCR of IDLV‐transduced and irradiated NHDF‐A at 1h, 4h, 8h, and 24h after irradiation did not identify more DSB sites than in non‐irradiated control cells. A possible reason for the low number of induced DSB sites is the presence of high numbers of non‐integrated, episomal IDLV vector copies. Thereby, an increased amplification of LTR‐LTR and LTR‐vector backbone sequences during LAM‐PCR that overlie the amplicons of the vector‐genome junctions is the result. Thus, the number of labeled DSB sites is assumed to be higher. In order to reduce the number of episomal amplicons, several strategies can be applied. One possibility would be to cut the isolated genomic DNA with restriction enzymes that specifically target the IDLV vector sequence. The resulting smaller vector fragments could subsequently be separated from the genomic DNA, which is then subjected to LAM‐PCR and deep sequencing. Furthermore, to increase the sensitivity during pyrosequencing, the number of reads for each sample can be increased. Here, each sample was sequenced at a depth of 800,000 reads. With the continued improvements in next‐generation sequencing technologies, additional increases in the sequencing depth should be possible, allowing deeper sequencing of vector‐genome junctions in the presence of vector amplicons.
In order to study the kinetics of DSB induction and repair after irradiation, two new approaches were developed and evaluated concerning their efficiency to label DSB sites in situ. The first approach was based on genome‐ wide labeling of DSB sites by the template‐independent DNA polymerase terminal deoxynucleotidyl transferase (Tdt), which adds nucleotides to free 3’ OH DNA termini in a random fashion [151]. In the second approach, free DNA ends were labeled by DNA ligases, which attach a specific DNA linker sequence to the DSB site. This method was termed LAM‐DST. In order to show that direct DSB labeling by Tdt and DNA ligases combined with LAM‐PCR can identify DSB, DSB sites were introduced into the genome by a CCR5‐specific ZFN. LAM‐PCR of Tdt‐labeled samples did not identify any of the reported ZFN on‐ or off‐target sites. Instead, unrelated genomic DNA sequences with an increased content of adenine and thymidine were amplified during LAM‐PCR. Since the PCR primer used during exponential amplification has a polyA tail, this implies that the primer binds unspecifically to DNA sequences that have a similar sequence to the Tdt‐mediated polyU tail. It may be possible that the transfection efficiency or ZFN activity was too low to induce efficient DSB levels for labeling and amplification. Therefore, Tdt‐based DSB labeling might still be useful to study on‐/off‐target ZFN activity and DSB repair kinetics despite unspecific DNA amplification. However, in order to use Tdt‐mediated DSB labeling for genome‐ wide DSB detection, the rate of unspecific DNA amplification needs to be greatly reduced and parameters for separating “true” DSB sites from unspecifically amplified sequences selected.
PCR on the CCR5 locus in samples treated with CCR ZFN and labeled by T4 DNA ligase or circLigase demonstrated that the ZFN was active and that the DNA ligases efficiently labeled the free DNA ends. However, subsequent LAM‐PCR on circLigase‐labeled samples did not identify the CCR5 ZFN on‐ and off‐target loci. This suggests that LAM‐DST can detect induced DSB, but that additional adaptations for the reduction of artifically induced DSB during processing and random amplification are required to enable genome‐wide DSB site identification. A potential source for artifically‐induced, non‐ZFN‐mediated DSB is the utilization of excessive concentrations of paraformaldehyde (PFA). PFA is used for the stabilization of the chromatin‐DNA complexes, and, if used at too high concentrations and for too long, it can result in the hydrolysis of the phosphodiester bonds in the DNA [152]. This creates free 3’ OH termini that are labeled by the DNA ligases. Secondly, linear double‐stranded DNA molecules are susceptible to destabilization, which can create free 3’ OH sites for ligation of 5’ phosphorylated DNA. This could be a potential source for the formation of DNA linker concatemers. Thus, in order to increase 97 DISCUSSION
the stability of DNA linker molecules against nuclease degradation, hairpin forming DNA sequences and phosphothioate residues at the 3’ DNA termini can be introduced [153].
Even though CCR5 ZFN‐induced DSB sites could not be efficiently identified by LAM‐DST and LAM‐PCR, a direct in situ ligation approach is a valuable tool to study DSB site distribution as well as designer nuclease activity sites. In contrast to classical methods used for studying DNA repair, in situ ligation offers several advantages. First, it enables direct marking of DSB sites at single‐nucleotide resolution, the analysis of the kinetics of DSB repair in different cell cycle phases and the influence of the genomic context on DSB distribution. Moreover, in situ DSB labeling is organism independent, thereby allowing the identification of DSB sites in multiple cell types and under various conditions. Further, it can help to explore the basis of genomic instability by combining data from in situ labeling with histone and transcriptional data sets [154]. The results gained from such studies could prove highly valuable for analyzing the mechanisms underlying cancer therapy resistance. Furthermore, analyzing the global distribution of endonuclease‐induced DSB sites such as CRISPRS/Cas9 and TALEN is important in order to assess and improve the safety of novel gene therapy treatments. Recent designer nuclease off‐target studies have shown the enormous potential of DSB detection sequencing methods, in line with the developments described here in this work [155‐157].
98 SUPPLEMENT
5. SUPPLEMENT
5.1 Supplementary Figures
A549 NHDF-A Photon Photon Ctr1Gy 4Gy Ctr 1Gy 4Gy
30min 30min
1h 1h
2h 2h
4h 4h
6h 6h
8h 8h
12h 12h
Supplementary Figure 1: Immunostaining of γH2AX foci in irradiated A549 and NHDF‐A
99 SUPPLEMENT
80 Non-irradiated A549 Photon 0.125 Gy Non-irradiated 0.25 Gy 0.5 Gy 1 Gy 0.125 Gy 60 2 Gy Cells [%] Cells 4 Gy
+ 0.25 Gy 10 Gy 0.5 Gy 40 1 Gy 2 Gy 20 4 Gy 10 Gy 0 0 5 10 15 20 25 30
100 Non-irradiated U87 Photon 0.125 Gy Non-irradiated 0.25 Gy 0.5 Gy 80 1 Gy 0.125 Gy 2 Gy
Cells [%] Frequencyof GFP 4 Gy
+ 10 Gy 0.25 Gy 60 0.5 Gy
40 1 Gy 2 Gy 20 4 Gy 10 Gy 0 0 5 10 15 20 25 30 35 100 NHDF-A Photon Non-irradiated Non-irradiated 1 Gy 4 Gy 80 1 Gy
Cells [%] Frequency of GFP 4 Gy + 60
40
20
Frequency ofGFP 0 0 5 10 15 20 25 30 35 Days after irradiation
Supplementary Figure 2: Frequency of EGFP+ cells transduced with IDLV during cultivation after irradiation.
20 -IR +Mirin -IR +NU7441 -IR +Mirin +NU7441 15 -IR Control
Cells [%] Cells 10 + EGFP 5
0 p3 p4 p6 p8
Supplementary Figure 3: Frequency of EGFP+ and IDLV‐transduced NHDF‐A with impaired NHEJ‐repair activity during expansion in cell culture. These cells were not exposed to irradiation. EGFP values were determined by flow cytometry.Control: active NHEJ‐repair activity; p: passage
100 SUPPLEMENT
7.0
6.0
5.0
4.0
3.0 HeLa Cells [%] + 2.0 GFP E 1.0
0.0 Untreated LD25 LD50 Doxorubicin
Supplementary Figure 4: Frequency of EGFP+, IDLV‐transduced HeLa cells treated with LD25 and LD50 doxorubicin after 21 days in culture. EGFP values were determined by flow cytometry.
25 6
5 20
4 15
H2AX Foci/Cell H2AX Foci/Cell 3 10 2
Average no 5 Average no 1
0 0 Untreated LD5 LD10 Untreated LD5 LD10 Doxorubicin Etoposide
Supplementary Figure 5: Average γH2AX foci number in NHDF‐A treated continusly with doxorubicin and etoposide for 43h. Error bars represent the standard deviation of the mean of the γH2AX foci number per cell.
101 SUPPLEMENT
16 14 A Photon D +IR Mirin 14 Proton 12 +IR Nu7441 Carbon 12 +IR M+N Natural 10 +IR Ctr 10 8 8 p4 6 6 A549 4 4 Frequency of DSB [%] Sites
2 [%] Sites DSB of Frequency 2
0 0 12345678910111213141516171819202122XY 12345678910111213141516171819202122XY Chromosome Chromosome B 16 14 Photon +IR Mirin 14 Proton 12 +IR Nu7441 Carbon 12 +IR M+N Natural 10 +IR Ctr 10 8 8 p6 6 6 PC3
NHDF-A 4 4 Frequency of DSBSites [%]
2 [%] Frequency of DSB Sites 2
0 0 12345678910111213141516171819202122XY 12345678910111213141516171819202122XY Chromosome Chromosome C 16 14 Photon +IR Mirin 14 Proton 12 +IR Nu7441 Carbon +IR M+N 12 10 Natural +IR Ctr 10 8 8 p8 6
U87 6 4 4 Frequency of DSB Sites [%] Sites DSB of Frequency
Frequency of DSB Sites [%] 2 2
0 0 12345678910111213141516171819202122XY 12345678910111213141516171819202122XY Chromosome Chromosome
Supplementary Figure 6: Chromosomal distribution of radiation‐induced DSB sites in A549 (A), PC3 (B), U87 (C) and NHDF‐A with impaired NHEJ‐repair activity (D).
102 SUPPLEMENT
In Gene In Gene
A 14 D 8 Photon +IR Mirin +IR Nu7441 12 Proton +IR M+N Carbon 6 +IR Ctr 10 Natural Random 8 p4 4 6 A549
4 2 Frequency of DSB Sites [%] Sites DSB of Frequency Frequency of DSB Sites [%] 2
0 0 10-5kb 5-0kb 0-10% 10-20% 20-30% 30-40% 40-50% 50-60% 60-70% 70-80% 80-90%90-100% 10-5kb 5-0kb 0-10% 10-20% 20-30% 30-40% 40-50% 50-60% 60-70% 70-80% 80-90%90-100% Upstream In Gene Upstream In Gene
B 14 8 Photon +IR Mirin +IR Nu7441 12 Proton +IR M+N Carbon 6 +IR Ctr 10 Natural Random 8 p6 4
PC3 6
4 2 NHDF-A Frequency of DSB Sites [%] Sites DSB of Frequency
Frequency of DSB [%] Sites 2
0 0 10-5kb 5-0kb 0-10% 10-20% 20-30% 30-40% 40-50% 50-60% 60-70% 70-80% 80-90%90-100% 10-5kb 5-0kb 0-10% 10-20% 20-30% 30-40% 40-50% 50-60% 60-70% 70-80% 80-90%90-100% Upstream In Gene Upstream In Gene
C 14 8 Photon +IR Mirin +IR Nu7441 12 Proton +IR M+N Carbon 6 +IR Ctr 10 Natural Random 8 p8 4
U87 6
4 2 Frequency of DSB Sites [%] Sites DSB of Frequency
Frequency of DSB Sites [%] 2
0 0 10-5kb 5-0kb 0-10% 10-20% 20-30% 30-40% 40-50% 50-60% 60-70% 70-80% 80-90%90-100% 10-5kb 5-0kb 0-10% 10-20% 20-30% 30-40% 40-50% 50-60% 60-70% 70-80% 80-90%90-100% Upstream In Gene Upstream In Gene
Supplementary Figure 7: Distribution of radiation‐induced DSB sites in genes and at the transcription start site (TSS). A549 (A), PC3 (B), U87 (C), NHDF‐A with impaired NHEJ‐repair activity (D).
Vector-genome junction Nucleotide Position in Sequence
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 A2420352223272626302831292930293130293130 T3645313938323834303231333333313131313131 G2017151417241820211918191920202020212019 C2017192422171821202120191918191819191819 Doxorubicin
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 A2428282629262929283030303125292929242426 T3232282524242732292529262532313032323430 G3024292927292622282425232424232423282424 Natural C1417152020211918152216212019171717161820
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 A3130293028292929303030302930293029303030 T2929292930293030303029292929303030303028 G2020212021212020202020202020212020202021 Random C2020222120212021212021212120212021212021
Supplementary Figure 8: Analysis of the genome sequence at doxorubicin‐induced DSB sites. Since the first two nucleotides frequently map to the LTR sequence of the IDLV, these nucleotides were not included for sequence analysis. In the random DSB data set (bottom), the first two nucleotides were included, since these sequences were generated in silico and do not contain LTR sequences.
103 SUPPLEMENT
AB60 100 IDLV
80 IR+IDLV ICLV 60
40 40 LTR Deletions [%] Deletions LTR 10 20
8 0 01234567891011121314151617181920 6 Size of LTR deletions [%]
20 [%] Deletions LTR LTR Deletions [%] 4 LTR Deletions [%] Deletions LTR
2
0 0 01234567891011121314151617181920 IDLV IR+IDLV IDLV Dox+IDLV ICLV Size of LTR deletions [%]
Irradiation Doxorubicin
C 80 IDLV IR+IDLV 60
40
20 LTR Deletions [%] Deletions LTR
0 Ctr M N MN Ctr M N MN Ctr M N MN passage 4 passage 6 passage 8
Supplementary Figure 9: Frequency of LTR deletions in integrated IDLV vector copies compared to an integrase‐competent lentivirus (ICLV).(A) Total frequency of IDLV and ICLV with deleted LTR sequences. (B) Size of LTR deletions in IDLV and ICLV. (C) LTR deletion frequency in NHDF‐A with impaired NHEJ‐repair activity
10.000,0
1.000,0
100,0 Signal Intensity 10,0
1,0 Photon Natural (IR) LD5 LD10 Natural (Dox) Random
Irradiation Doxorubicin
Supplementary Figure 10: The signal intensity of the DNaseI HS does not influence distribution of radiation‐ and doxorubicin‐ induced DSB sites. Boxes represent upper and lower quantiles, and middle line indicates the mean.
104 SUPPLEMENT
3.5 Photon Proton 3.0 Carbon Natural 2.5
2.0 A549 1.5
1.0
0.5 Enrichment of DSB at DHS compared to to DHSRandomat DSB compared Enrichmentof
0.0 0246810 Radius around DSB Site [kb]
3.5 Natural (IR) Photon 3.0 LD5 LD10 2.5 Natural (TOP2)
2.0 NHDF-A 1.5
1.0
0.5 Enrichment of DSB at DHS compared to Random
0.0 0246810 Radius around DSB Site [kb]
Supplementary Figure 11: DSB sites do not cluster in proximity to DnaseI HS, but are enriched inside DnaseI HS. The red line indicates the frequency for the random DSB data set.
Irradiated NHDF-A Control +Mirin +Nu7441 +Mirin +Nu7441 1.3
1.0 Euchromatin 0.7 1.3
1.0
Heterochromatin 0.7 1.3
1.0 Border Region 0.7 p4 p6 p8 p4 p6 p8 p4 p6 p8 p4 p6 p8
Supplementary Figure 12: Distribution of radiation‐induced DSB sites in eu‐ and heterochromatin of NHEJ‐impaired NHDF‐A. Similar to NHEJ‐proficient cells, the DSB frequency increases in border regions and euchromatin in passage 4, 6, and 8, whereas the DSB frequency in heterochromatin decreases during cultivation. Numbers on y‐axis represent enrichment values of DSB frequency compared to random DSB frequency.
105 SUPPLEMENT
Irradiated NHDF-A Control +Mirin +Nu7441 +Mirin +Nu7441 1.5
H3K4me1 1.0 0.8 1.5 H3K4me2 1.0 0.8 1.5 H3K4me3 1.0 0.8 1.5
H3K9ac 1.0 0.8 1.5
H3K27ac 1.0 0.8 1.5 H3K36me3 1.0 0.8 1.5 H4K20me1 1.0 0.8 1.5 H3K9me3 1.0 0.8 1.5 H3K27me3 1.0 0.8 1.5 H3K79me2 1.0 0.8 p4 p6 p8 p4 p6 p8 p4 p6 p8 p4 p6 p8
Supplementary Figure 13: Histone modifications associated to radiation‐induced DSB sites in NHEJ‐impaired NHDF‐A in three passages. Numbers on y‐axis represent enrichment values of histone frequency compared to histone frequency at randomly‐ distributed DSB.
106 SUPPLEMENT
+IR Control +IR +Mirin
H3K4me1 H3K4me2 H3K4me3 H3K9ac H3K27ac H3K36me3 H4K20me1 H3K9me3 H3K27me3 H3K79me2 H3K4me1 H3K4me2 H3K4me3 H3K9ac H3K27ac H3K36me3 H4K20me1 H3K9me3 H3K27me3 H3K79me2 H3K4me1 1,091,101,211,151,151,131,111,331,14 H3K4me1 1,121,151,291,391,301,281,081,261,21 H3K4me2 1,031,321,291,181,151,211,491,20 H3K4me2 0,780,961,060,980,930,891,060,93 H3K4me3 1,021,050,971,001,031,250,97 H3K4me3 0,730,890,750,710,720,870,75 H3K9ac 0,99 1,02 1,09 1,04 1,38 1,04 H3K9ac 0,86 0,86 0,83 0,77 1,10 0,84 H3K27ac 1,02 1,12 1,25 1,55 1,09 H3K27ac 0,96 1,12 0,99 1,34 1,07 p4 H3K36me3 1,02 1,38 1,33 1,22 p4 H3K36me3 0,82 1,15 1,10 1,00 H4K20me1 1,13 1,97 1,64 H4K20me1 0,86 1,39 1,31 H3K9me3 1,15 2,06 H3K9me3 0,75 1,37 H3K27me3 1,26 H3K27me3 0,84 H3K79me2 H3K79me2
H3K4me1 H3K4me2 H3K4me3 H3K9ac H3K27ac H3K36me3 H4K20me1 H3K9me3 H3K27me3 H3K79me2 H3K4me1 H3K4me2 H3K4me3 H3K9ac H3K27ac H3K36me3 H4K20me1 H3K9me3 H3K27me3 H3K79me2 H3K4me1 1,041,081,221,221,191,131,151,251,21 H3K4me1 1,211,161,391,331,331,291,291,491,35 H3K4me2 1,041,321,301,251,271,151,441,29 H3K4me2 0,951,311,241,201,181,131,481,20 H3K4me3 1,011,110,961,031,041,260,99 H3K4me3 0,940,980,900,900,861,140,91 H3K9ac 1,08 1,05 1,14 1,07 1,46 1,09 H3K9ac 1,06 1,10 1,09 1,12 1,48 1,12 H3K27ac 1,08 1,18 1,21 1,63 1,18 H3K27ac 1,08 1,12 1,25 1,78 1,16 H3K36me3 1,11 1,42 1,53 1,26 p6 H3K36me3 1,00 1,33 1,47 1,20 p6 H4K20me1 1,03 1,96 1,85 H4K20me1 1,10 1,95 1,62 H3K9me3 1,18 2,00 H3K9me3 1,09 2,11 H3K27me3 1,23 H3K27me3 1,29 H3K79me2 H3K79me2
H3K4me1 H3K4me2 H3K4me3 H3K9ac H3K27ac H3K36me3 H4K20me1 H3K9me3 H3K27me3 H3K79me2 H3K4me1 H3K4me2 H3K4me3 H3K9ac H3K27ac H3K36me3 H4K20me1 H3K9me3 H3K27me3 H3K79me2 H3K4me1 1,231,191,481,471,381,551,321,471,47 H3K4me1 1,071,071,241,231,191,191,161,261,21 H3K4me2 1,041,501,501,281,541,241,461,43 H3K4me2 0,800,981,000,940,970,971,060,96 H3K4me3 1,131,160,961,250,951,101,09 H3K4me3 0,730,810,740,810,770,870,78 H3K9ac 1,27 1,19 1,45 1,28 1,66 1,30 H3K9ac 0,82 0,86 0,89 0,95 1,12 0,88 H3K27ac 1,22 1,55 1,40 1,93 1,42 H3K27ac 0,88 0,96 1,02 1,23 0,96 p8 H3K36me3 1,22 1,43 1,46 1,41 p8 H3K36me3 0,85 1,14 1,11 1,01 H4K20me1 1,41 2,44 2,15 H4K20me1 0,92 1,55 1,36 H3K9me3 1,11 2,53 H3K9me3 0,85 1,74 H3K27me3 1,48 H3K27me3 0,94 H3K79me2 H3K79me2
+IR +Nu7441 +IR +Mirin +Nu7441 H3K4me1 H3K4me2 H3K4me3 H3K9ac H3K27ac H3K36me3 H4K20me1 H3K9me3 H3K27me3 H3K79me2 H3K4me1 H3K4me2 H3K4me3 H3K9ac H3K27ac H3K36me3 H4K20me1 H3K9me3 H3K27me3 H3K79me2 H3K4me1 1,111,101,251,301,231,241,171,341,29 H3K4me1 1,051,051,221,151,101,131,001,281,10 H3K4me2 0,881,181,221,071,111,111,171,14 H3K4me2 0,971,311,211,111,181,041,371,15 H3K4me3 0,850,940,780,870,901,000,89 H3K4me3 1,021,060,911,000,911,110,98 H3K9ac 0,98 0,91 0,98 0,99 1,28 1,01 H3K9ac 1,03 0,96 1,02 1,02 1,36 0,99 H3K27ac 0,97 1,10 1,24 1,41 1,15 H3K27ac 1,01 1,12 1,06 1,50 1,11 p4 H3K36me3 0,95 1,23 1,07 1,09 p4 H3K36me3 0,99 1,21 1,19 1,15 H4K20me1 1,04 1,72 1,59 H4K20me1 1,02 1,79 1,55 H3K9me3 0,86 1,87 H3K9me3 0,99 1,78 H3K27me3 0,95 H3K27me3 1,04 H3K79me2 H3K79me2
H3K4me1 H3K4me2 H3K4me3 H3K9ac H3K27ac H3K36me3 H4K20me1 H3K9me3 H3K27me3 H3K79me2 H3K4me1 H3K4me2 H3K4me3 H3K9ac H3K27ac H3K36me3 H4K20me1 H3K9me3 H3K27me3 H3K79me2 H3K4me1 1,101,071,291,311,261,261,261,311,29 H3K4me1 1,121,171,301,301,231,271,301,361,23 H3K4me2 0,800,981,010,950,930,900,980,94 H3K4me2 1,051,321,361,171,241,271,441,26 H3K4me3 0,720,770,700,710,690,760,73 H3K4me3 1,061,160,951,041,151,171,03 H3K9ac 0,82 0,86 0,87 0,87 1,13 0,86 H3K9ac 1,10 1,05 1,17 1,22 1,52 1,12 H3K27ac 0,92 0,99 1,04 1,16 1,01 H3K27ac 1,08 1,25 1,39 1,83 1,25 H3K36me3 0,81 1,01 0,93 0,95 p6 H3K36me3 1,05 1,46 1,43 1,27 p6 H4K20me1 0,96 1,41 1,36 H4K20me1 1,25 2,06 1,80 H3K9me3 0,81 1,66 H3K9me3 1,25 2,34 H3K27me3 0,79 H3K27me3 1,42 H3K79me2 H3K79me2
H3K4me1 H3K4me2 H3K4me3 H3K9ac H3K27ac H3K36me3 H4K20me1 H3K9me3 H3K27me3 H3K79me2 H3K4me1 H3K4me2 H3K4me3 H3K9ac H3K27ac H3K36me3 H4K20me1 H3K9me3 H3K27me3 H3K79me2 H3K4me1 1,221,171,501,461,381,551,361,501,45 H3K4me1 1,091,081,251,251,231,181,131,311,19 H3K4me2 0,790,990,990,970,950,951,060,92 H3K4me2 1,031,341,341,251,231,211,461,23 H3K4me3 0,680,730,680,690,650,690,65 H3K4me3 1,051,110,991,041,011,250,99 H3K9ac 0,89 0,95 0,97 1,00 1,30 0,90 H3K9ac 1,09 1,10 1,12 1,10 1,45 1,11 H3K27ac 0,98 1,06 1,12 1,40 1,01 H3K27ac 1,13 1,21 1,25 1,66 1,23 p8 H3K36me3 0,79 0,99 0,95 0,89 p8 H3K36me3 1,11 1,43 1,41 1,31 H4K20me1 1,15 1,92 1,56 H4K20me1 1,22 2,03 1,76 H3K9me3 0,70 1,30 H3K9me3 1,12 2,22 H3K27me3 0,72 H3K27me3 1,30 H3K79me2 H3K79me2
Supplementary Figure 14: Radiation‐induced and naturally‐occurring DSB in NHEJ‐impaired NHDF‐A cells show association to genomic regions marked by H3K27me3 in combination with H3K4me3, H3K9ac, H3K27ac and H3K36me3. Numbers indicate the enrichment in the histone‐per‐DSB frequency compared to the random DSB data set. IR: irradiated, p: passage
H3K4me1 H3K4me2 H3K4me3 H3K9ac H3K27ac H3K36me3 H4K20me1 H3K9me3 H3K27me3 H3K4me1 1,061,141,071,141,111,091,031,29 H3K4me2 1,011,061,111,020,991,121,24 H3K4me3 0,991,021,071,131,471,58 H3K9ac 1,06 1,10 1,18 1,36 1,71 Natural H3K27ac 1,13 1,13 1,34 1,66 H3K36me3 1,03 1,14 1,36 H4K20me1 1,03 1,18 H3K9me3 1,03
H3K4me1 H3K4me2 H3K4me3 H3K9ac H3K27ac H3K36me3 H4K20me1 H3K9me3 H3K27me3 H3K4me1 1,051,061,071,071,051,061,101,12 H3K4me2 1,011,031,041,071,121,091,11 H3K4me3 1,021,001,041,141,161,25 H3K9ac 1,04 1,04 1,11 1,15 1,36 Photon H3K27ac 1,06 1,10 1,21 1,38 H3K36me3 1,06 1,15 1,21 H4K20me1 1,07 1,15 H3K9me3 1,01
H3K4me1 H3K4me2 H3K4me3 H3K9ac H3K27ac H3K36me3 H4K20me1 H3K9me3 H3K27me3 H3K4me1 1,051,101,111,101,011,031,091,17 H3K4me2 1,081,101,091,060,961,041,04 H3K4me3 1,041,031,061,111,121,21 H3K9ac 1,05 1,10 1,07 1,18 1,30 Proton H3K27ac 1,06 1,04 1,18 1,39 H3K36me3 1,01 1,16 1,11 H4K20me1 1,06 1,14 H3K9me3 1,00
H3K4me1 H3K4me2 H3K4me3 H3K9ac H3K27ac H3K36me3 H4K20me1 H3K9me3 H3K27me3 H3K4me1 1,021,031,021,061,151,131,131,11 H3K4me2 1,030,991,031,101,101,171,16 H3K4me3 0,991,071,121,171,281,28 H3K9ac 1,07 1,04 1,11 1,22 1,70 Carbon H3K27ac 1,10 1,19 1,34 1,86 H3K36me3 1,04 1,16 1,25 H4K20me1 1,03 1,21 H3K9me3 1,03
Supplementary Figure 15: Radiation‐induced and naturally‐occurring DSB in A549 cells show association to genomic regions marked by H3K27me3 in combination with H3K4me3, H3K9ac, H3K27ac and H3K36me3. Numbers indicate the enrichment in the histone per DSB frequency compared to the random DSB data set. The radiation source is indicated on the left.
107 SUPPLEMENT
A 4.5 Radiation 4.0 Natural 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 Enrichment Factor Compared to Random to Compared Factor Enrichment
B 4.5 Doxorubicin 4.0 Natural 3.5 3.0 2.5 2.0 1.5 1.0 0.5
Enrichment Factor Compared0.0 toRandom
Supplementary Figure 16: Upstream regulators of IDLV‐marked genes in (A) irradiated and (B) doxorubicin‐treated NHDF‐A
108 SUPPLEMENT
100 A Nat-Occurring Overlapping 80 IR-related
60
40
DSB Hotspot DSBFrequency Hotspot [%] 20
0 p4 p6 p8
70 B Nat‐Occurring 60 Overlapping IR‐related 50
40
30
20 Euchromatin DSB Hotspot Frequency [%] 10
0 70 Nat‐Occurring 60 Overlapping IR‐related 50
40
30
20 DSB Hotspot Frequency [%]
Heterochromatin 10
0 70 Nat‐Occurring
60 Overlapping IR‐related 50
40
30
20 Border RegionBorder DSB Hotspot Frequency [%] 10
0 p4 p6 p8
Supplementary Figure 17: Kinetics and distribution of DSB hotspots with respect to chromatin regions. (A) The frequency of naturally‐occurring DSB hotspots decreased during expansion, whereas the frequency of radiation‐related DSB hotspots increased. (B) Naturally‐occurring DSB hotspots decreased in euchromatin, heterochromatin and in border regions. In contreast, the frequency of overlapping DBS hotspots remained constant and radiation‐related hotspots increased.
Position in genome downstream of vector-genome junction 1234567891011121314151617181920 A7961292332373536353434333234333535323232 T 7 19 54 57 33 33 31 30 30 31 25 29 30 34 31 30 29 32 27 28 G 7 11 8 9 15 15 16 14 19 21 23 18 19 18 21 19 20 20 24 19
Nucleotide C 7 9 8 11 20 15 18 19 16 14 17 19 18 14 15 17 16 15 17 20
Supplementary Figure 18: Nucleotide frequency at indicated positions downstream of vector‐genome junction. The first two nucleotides (position 1 and 2) have a high frequency of the adenosine, which is most likely to belong to the polyU sequence added to the free DNA ends by Tdt enzyme. Thus, the first nucleotide of the genomic sequence is assumed to be at position 3.
109 SUPPLEMENT
Supplementary Figure 19: CCR5 genomic locus with ligated ssGRVU5A DNA linker. Primers used during locus‐specific PCR and LAM‐PCR are indicated. The blue triangle marks the CCR5 ZFN target site.
110 SUPPLEMENT
5.2 SupplementaryTables
Supplementary Table 1: Number of unique DSB in cancer cell lines and primary human fibroblasts.
Cell Types DSB Induction Source Unique DSB sites Irradiation Photon 2540 Irradiation Proton 1102 A549 Irradiation Carbon 825 Natural 188 Irradiation Photon 642 Irradiation Proton 588 PC3 Irradiation Carbon 330 Natural 88 Irradiation Photon 1261 Irradiation Proton 510 U87 Irradiation Carbon 330 Natural 124 Irradiation Photon 6959 NHDF‐A Natural 3887 Chemotherapy LD5 Doxorubicin 4003 NHDF‐A Chemotherapy LD10 Doxorubicin 3725 Natural 263 Random ‐ 4742
Cell Type DSB Induction NHEJ‐repair Blocking Agent Passage Unique DSB sites Irradiation +Mirin p4 764 Irradiation +Mirin p6 1459 Irradiation +Mirin p8 3929 Irradiation +Nu7441 p4 616 Irradiation +Nu7441 p6 1927 Irradiation +Nu7441 p8 3131 Irradiation +Mirin +Nu7441 p4 787 Irradiation +Mirin +Nu7441 p6 2151 Irradiation +Mirin +Nu7441 p8 4243 Irradiation Control p4 803 Irradiation Control p6 1194 Irradiation Control p8 1759 NHDF‐A +Mirin p4 975 +Mirin p6 488 +Mirin p8 1036 +Nu7441 p4 897 +Nu7441 p6 326 +Nu7441 p8 1710 +Mirin +Nu7441 p4 735 +Mirin +Nu7441 p6 393 +Mirin +Nu7441 p8 667 Control p4 610 Control p6 453 Control p8 329
111 SUPPLEMENT
Supplementary Table 2: Frequency of DSB sites in RefSeq genes, at transcription start site and CpG islands. Integration site data for integrase‐competent lentiviruses (ICLV) were obtained from [88]. No values for DSB frequency in NHDF‐A at CpG islands were available, since the function was deactivated in HISAP. DSB Frequency [%] In RefSeq At TSS At CpG islands Cell Types DSB Induction Source Genes [+10kb] [≤1kb] Irradiation Photon 43.9 5.2 2.8 Irradiation Proton 47.9 6.1 5.9 A549 Irradiation Carbon 40.5 4.2 3.0 Natural 53.2 3.7 5.9 Irradiation Photon 45.0 5.6 1.8 Irradiation Proton 38.1 5.1 4.1 PC3 Irradiation Carbon 40.9 3.6 1.7 Natural 36.4 9.1 3.4 Irradiation Photon 41.5 4.8 2.1 Irradiation Proton 43.8 5.5 3.3 U87 Irradiation Carbon 43.0 3.9 1.8 Natural 48.0 2.4 10.5 Irradiation Photon 51.7 5.7 ‐ NHDF‐A Natural 51.0 6.9 ‐ Chemotherapy LD5 Doxorubicin 65.5 4.6 ‐ NHDF‐A Chemotherapy LD10 Doxorubicin 65.0 4.4 ‐ Natural 45.7 4.1 ‐ Random ‐ ‐ 40.2 4.7 2.6 ICLV ‐ ‐ 74.6 2.2 1.6
Supplementary Table 3: Frequency of lentiviral ‘GT’ dinucleotide at vector‐genome junctions. #: Number of sequences with GT dinucleotide at vector‐genome junction. # ‘GT’ p‐value (Fisher’s Cell Types DSB induction Source ‘GT’ Frequency Dinucleotide exact test) Irradiation Photon 2090 30.0 1.14E‐246 NHDF‐A Natural 2364 60.8 0.00E+00 Chemotherapy LD5 Doxorubicin 2560 64.0 0.00E+00 NHDF‐A Chemotherapy LD10 Doxorubicin 2394 64.3 0.00E+00 Natural 132 50.2 4.80E‐168 Random 230 4.9
112 SUPPLEMENT
Supplementary Table 4: Frequency of DSB sites located inside DNaseI HS.
DSB Enrichment compared to p‐value (Fisher’s Cell Types DSB induction Source Frequency Random exact test) Irradiation Photon 3.43 1.71 3.12E‐04 Irradiation Proton 3.90 1.95 2.82E‐04 A549 Irradiation Carbon 4.00 2.00 6.09E‐04 Natural 5.85 2.92 5.99E‐04 Random (A549) 2.00 ‐‐ Irradiation Photon 1.87 1.00 9.73E‐01 NHDF‐A Natural 1.03 0.55 1.24E‐03 Chemotherapy LD5 4.42 2.35 5.45E‐12 NHDF‐A Chemotherapy LD10 4.40 2.34 1.36E‐11 Natural 5.24 2.79 1.62E‐04 Random (NHDF‐A) 1.88 ‐‐
Enrichment p‐value NHEJ‐repair DSB Cell Type DSB induction Passage compared to (Fisher’s Blocking Agent Frequency Random exact test) Irradiation p4 +Mirin 1.44 0.77 4.01E‐01 Irradiation p6 +Mirin 1.23 0.66 9.90E‐02 Irradiation p8 +Mirin 1.86 0.99 9.49E‐01 Irradiation p4 +Nu7441 2.92 1.56 8.10E‐02 Irradiation p6 +Nu7441 1.40 0.75 1.78E‐01 Irradiation p8 +Nu7441 1.31 0.70 5.30E‐02 NHDF‐A Irradiation p4 +Mirin +Nu7441 1.78 0.95 8.51E‐01 Irradiation p6 +Mirin +Nu7441 1.16 0.62 3.10E‐01 Irradiation p8 +Mirin +Nu7441 1.46 0.78 1.26E‐01 Irradiation p4 Control 2.86 1.53 6.60E‐02 Irradiation p6 Control 1.34 0.71 2.08E‐01 Irradiation p8 Control 1.42 0.76 2.14E‐01 p4 +Mirin 1.85 0.98 9.49E‐01 p6 +Mirin 1.64 0.87 7.11E‐01 p8 +Mirin 2.41 1.29 2.61E‐01 p4 +Nu7441 2.01 1.07 7.94E‐01 p6 +Nu7441 1.84 0.98 9.63E‐01 p8 +Nu7441 2.22 1.18 3.78E‐01 NHDF‐A p4 +Mirin +Nu7441 1.50 0.80 4.74E‐01 p6 +Mirin +Nu7441 2.29 1.22 5.65E‐01 p8 +Mirin +Nu7441 1.65 0.88 6.83E‐01 p4 Control 2.13 1.14 6.65E‐01 p6 Control 4.19 2.23 1.00E‐03 p8 Control 1.13 1.13 7.47E‐01 Random (NHDF‐A) 1.88 ‐ ‐
113 SUPPLEMENT
Supplementary Table 5: Frequency of DSB sites located in border regions, eu‐ and heterochromatin
Cell Type DSB Induction Source Euchromatin Heterochromatin Border Regions Irradiation Photon 28.1 29.8 34.0 Irradiation Proton 31.9 26.7 35.9 A549 Irradiation Carbon 26.3 32.0 32.2 Natural 29.8 22.3 41.0 Random (A549) 24.2 34.8 30.5 Irradiation Photon 18.0 29.4 37.3 Natural 18.0 26.4 33.4 NHDF‐A Chemotherapy LD5 Doxorubicin 17.1 9.7 47.6 Chemotherapy LD10 Doxorubicin 16.1 10.6 47.2 Natural 11.2 29.6 39.0 Random (NHDF‐A) 15.2 34.3 34.0
Cell DSB NHEJ‐repair Border Type induction Passage Blocking Agent Euchromatin Heterochromatin Region Irradiation p4 +Mirin 11.8 29.5 34.7 Irradiation p6 +Mirin 12.3 28.2 33.7 Irradiation p8 +Mirin 14.7 29.1 35.2 Irradiation p4 +Nu7441 13.0 30.5 35.2 Irradiation p6 +Nu7441 13.1 30.7 31.3 NHDF‐ Irradiation p8 +Nu7441 12.7 28.5 38.5 A Irradiation p4 +Mirin +Nu7441 13.7 31.5 32.1 Irradiation p6 +Mirin +Nu7441 13.7 29.3 31.0 Irradiation p8 +Mirin +Nu7441 14.3 29.9 35.3 Irradiation p4 Control 13.7 34.1 33.7 Irradiation p6 Control 13.4 29.6 30.7 Irradiation p8 Control 12.7 29.8 36.6 p4 +Mirin 13.8 35.8 32.5 p6 +Mirin 16.4 28.1 35.5 p8 +Mirin 16.8 32.1 33.2 p4 +Nu7441 11.7 37.1 34.3 p6 +Nu7441 13.8 28.2 40.2 NHDF‐ p8 +Nu7441 18.2 26.4 38.5 A p4 +Mirin +Nu7441 12.5 31.7 35.5 p6 +Mirin +Nu7441 15.5 28.0 36.9 p8 +Mirin +Nu7441 13.3 31.9 32.8 p4 Control 13.4 33.3 34.6 p6 Control 14.6 36.4 32.9 p8 Control 11.2 35.0 31.3 Random (NHDF‐A) 15.2 34.3 34.0
114 SUPPLEMENT
Supplementary Table 6: Frequency and enrichment of DSB at CTCF Binding sites compared to random DSB data set.
DSB Enrichment p‐value Cell Types DSB induction Source Frequency in compared to (Fisher’s exact CTCF BS Random test) Irradiation Photon 10.8 1.4 2.7E‐04 Irradiation Proton 10.7 1.3 8.6E‐03 A549 Irradiation Carbon 13.6 1.7 3.0E‐06 Natural 15.4 1.9 1.3E‐03 Random (A549) 8.0 ‐ Irradiation Photon 19.5 1.1 3.5E‐06 NHDF‐A Natural 19.6 1.1 2.9E‐03 Chemotherapy LD5 Doxorubicin 32.1 1.8 1.6E‐57 NHDF‐A Chemotherapy LD10 Doxorubicin 32.6 1.9 7.2E‐59 Natural 24.0 1.4 7.0E‐03 Random (NHDF‐A) 17.4 ‐‐
Enrichment p‐value DSB NHEJ‐repair DSB Frequency Cell Type Passage compared to (Fisher’s induction Blocking Agent in CTCF BS Random exact test) Irradiation p4 +Mirin 17.0 1.0 7.9E‐01 Irradiation p6 +Mirin 17.0 1.0 7.1E‐01 Irradiation p8 +Mirin 18.0 1.0 4.8E‐01 Irradiation p4 +Nu7441 18.3 1.1 5.7E‐01 Irradiation p6 +Nu7441 18.2 1.0 4.7E‐01 Irradiation p8 +Nu7441 22.9 1.3 2.5E‐09 NHDF‐A Irradiation p4 +Mirin +Nu7441 14.7 0.9 6.4E‐02 Irradiation p6 +Mirin +Nu7441 15.7 0.9 8.0E‐02 Irradiation p8 +Mirin +Nu7441 17.7 1.0 7.3E‐01 Irradiation p4 Control 18.8 1.1 3.4E‐01 Irradiation p6 Control 15.9 0.9 2.2E‐01 Irradiation p8 Control 18.9 1.1 1.6E‐01 p4 +Mirin 17.2 1.0 8.9E‐01 p6 +Mirin 20.9 1.2 5.5E‐02 p8 +Mirin 16.7 1.0 5.8E‐01 p4 +Nu7441 15.9 0.9 2.8E‐01 p6 +Nu7441 20.2 1.2 2.0E‐01 p8 +Nu7441 19.8 1.1 2.7E‐02 NHDF‐A p4 +Mirin +Nu7441 17.8 1.0 7.9E‐01 p6 +Mirin +Nu7441 17.3 1.0 9.5E‐01 p8 +Mirin +Nu7441 16.6 1.0 6.2E‐01 p4 Control 19.2 1.1 2.8E‐01 p6 Control 19.4 1.1 2.8E‐01 p8 Control 15.8 0.9 4.6E‐01 Random 17.4 ‐ ‐
115 SUPPLEMENT
Supplementary Table 7: Average number of histone modifications per DSB in A549 and NHDF‐A
A549 Random Histone Modification Photon Proton Carbon Natural H3K4me1 0.98 1.34 0.91 1.39 0.80 H3K4me2 0.60 0.92 0.63 0.99 0.50 H3K4me3 0.56 0.71 0.57 0.74 0.45 H3K9ac 0.22 0.29 0.26 0.35 0.18 H3K27ac 0.41 0.58 0.46 0.74 0.32 H3K36me3 0.85 1.13 0.80 1.06 0.66 H4K20me1 0.44 0.70 0.44 0.58 0.36 H3K9me3 0.50 0.38 0.51 0.38 0.48 H3K27me3 0.82 0.66 0.91 0.65 0.67
NHDF‐A Random Histone Photon Natural LD5 Doxo LD10 Doxo Natural Modification H3K4me1 1.14 1.93 1.01 1.05 0.60 0.46 H3K4me2 1.19 1.92 0.67 0.73 0.37 0.31 H3K4me3 1.09 1.71 0.49 0.53 0.32 0.21 H3K9ac 1.01 1.77 0.86 0.86 0.45 0.33 H3K27ac 1.17 1.99 0.91 0.96 0.45 0.33 H3K36me3 1.28 2.26 0.96 0.95 0.55 0.35 H4K20me1 1.41 2.16 0.63 0.69 0.43 0.27 H3K9me3 0.83 1.39 0.35 0.31 0.45 0.43 H3K27me3 1.19 1.75 0.34 0.37 0.47 0.44 H3K79me2 1.38 2.33 0.70 0.74 0.38 0.22
NHDF‐A +IR Ctr +IR M +IR N +IR MN Histone p4 p6 p8 p4 p6 p8 p4 p6 p8 p4 p6 p8 Modification H3K4me1 1.59 1.50 1.85 1.66 1.80 1.57 1.65 1.59 1.82 1.51 1.68 1.56 H3K4me2 1.46 1.43 1.51 1.34 1.55 1.37 1.41 1.50 1.61 1.32 1.43 1.43 H3K4me3 1.31 1.31 1.22 1.32 1.28 1.24 1.27 1.32 1.31 1.28 1.35 1.28 H3K9ac 1.50 1.49 1.81 1.52 1.75 1.54 1.58 1.64 1.91 1.54 1.58 1.56 H3K27ac 1.38 1.53 1.76 1.69 1.60 1.49 1.63 1.69 1.88 1.39 1.62 1.55 H3K36me3 1.20 1.25 1.28 1.29 1.29 1.24 1.16 1.29 1.32 1.13 1.17 1.27 H4K20me1 1.15 1.25 1.49 1.21 1.26 1.20 1.20 1.31 1.68 1.12 1.24 1.21 H3K9me3 1.17 1.23 1.17 1.16 1.22 1.21 1.18 1.35 1.29 1.07 1.30 1.23 H3K27me3 1.31 1.36 1.31 1.18 1.40 1.30 1.21 1.28 1.31 1.27 1.31 1.37 H3K79me2 1.27 1.39 1.48 1.26 1.36 1.33 1.50 1.43 1.57 1.16 1.37 1.35
116 SUPPLEMENT
NHDF‐A ‐IR Ctr ‐IR M ‐IR N ‐IR MN Histone p4 p6 p8 p4 p6 p8 p4 p6 p8 p4 p6 p8 Modification H3K4me1 1.59 1.66 1.59 1.42 1.84 1.52 1.56 1.89 1.51 1.47 1.64 1.54 H3K4me2 1.30 1.46 1.49 1.35 1.43 1.25 1.35 1.60 1.35 1.32 1.37 1.26 H3K4me3 1.19 1.35 1.23 1.30 1.56 1.21 1.21 1.37 1.26 1.12 1.35 1.21 H3K9ac 1.53 1.61 1.50 1.46 1.74 1.41 1.38 1.73 1.49 1.27 1.45 1.36 H3K27ac 1.50 1.28 1.30 1.38 1.90 1.36 1.48 1.61 1.56 1.36 1.41 1.26 H3K36me3 1.22 1.20 1.17 1.14 1.22 1.23 1.13 1.20 1.17 1.19 1.22 1.21 H4K20me1 1.31 1.12 1.11 1.18 1.74 1.24 1.13 1.35 1.17 1.26 1.25 1.18 H3K9me3 1.13 1.17 1.07 1.26 1.33 1.11 1.17 1.23 1.18 1.11 1.38 1.15 H3K27me3 1.17 1.18 1.36 1.24 1.35 1.28 1.29 1.34 1.27 1.25 1.13 1.26 H3K79me2 1.29 1.23 1.28 1.27 1.68 1.40 1.17 1.44 1.29 1.34 1.26 1.21
Supplementary Table 8: Upstream Regulators associated to genes that show intragenic labeling of radiation‐induced DSB
Molecule Type Upstream Regulator Target Molecules AFF1 EPHA7,PTEN AJUBA DOCK1,PTK2 DNMT3L IGF2R,SNRPN AR,ARHGAP4,BCR,CAMK1D,CDC42BPB,CYFIP1,DIAPH2,DIAPH3, DIP2A,DOCK1,ELMO1,ETS1,EXT1,FLT1,FYN,GPATCH2L,MAP3K5, ERG MYO10,NLGN1,NRCAM,NRG1,PHACTR2,PLAUR,POU2F1,RGS3,R PS11,RSU1,SVIL,TSC22D3,UTRN,WIPF1 ABCC1,ADAM12,AGPS,ALDH1A3,ANK3,ASAP1,B4GALT1,BCL9,C ASZ1,COBLL1,DDX3X,DLG1,DOCK1,ELOVL6,FLRT2,GAS2,HIPK3,L
FOS AMA3,LARGE,MAP4K4,MLH1,PFN1,PRKCE,PTPRD,QKI,RARA,RIP K1,RXRA,SDK1,SMAD7,SULF2,TAOK1,UTRN,ZHX2 Regulator
GABPA AURKA,FAS,SKP2,SPI1 ACAN,ATG5,COL1A2,DCLRE1C,FAS,FDXR,LGR6,LIG1,PRIM2,SFRP HDAC2 1,TP63 CACNA2D1,CHRDL1,CNKSR2,CNTN1,DGKB,DLG2,GABRA1,GPHN Transcription HDAC4 ,GPM6A,MAPK10,NRCAM,NTM,PRKCA,PRKCB,RAB3C,RGS7 HOXC6 ABCB1,BMP7,CNTN1,COL23A1,PDE4DIP,SIPA1L1,ZNF536 HR ANO1,COL25A1,CSNK2A2,CTNND2,DYM,MAGI2 ABCB1,BMP7,CDK1,FBLN1,FGF1,HERC4,JARID2,JMJD6,MLH1,M HSF1 SH4,NFATC2,PPP3CB,RAD51B,RAD51D,RARRES3,STAG2,STIP1,T CP1,WNT3 AQP1,CCND2,DOCK1,DOCK9,IFI16,LHFP,LRP1, IKZF1 NCAM1,NOTCH3,PRKCQ,SPI1,SULF2,VWF JARID2 CCND2,PTEN 117 SUPPLEMENT
ABCB1,BCL9,CDK1,COL1A2,HNRNPA2B1,IFI16,IGF2R,ITGAV,PLA JUN UR,PRKCE,PTEN,PTPRD,RARA,RXRA,SDK1,SMAD7,SNRPN,STAT1, SULF2,XYLT1 KLF3 ANK1,KLF12,SELL MEOX2 ITGAV,LRP1 ACTG1,CTNNA1,ETS1,ILF3,LPIN2,MAL,MEF2A,NR2C2,PRKCB,RIP SATB1 K1,SELL,SIPA1L2,SPI1,TAF4B,TAOK1,TMEM117,TNFRSF8,TSC22 D3,VTA1 COL4A1,COL4A2,COL4A5,COL4A6,COL5A1,CYFIP1,ITGA5,PLAUR SPDEF ,PRKCA,SMAD4 TAF7L AR,CPA6,FSCN1,HK1,LPIN1,NR6A1,TSKS,ZFX ABCB1,ABCC1,APBB2,AR,ASTN2,ATXN1,AURKA,BAI3,BAK1,BCAS 3,BRAF,C12orf5,CAMK2B,CARD11,CCND2,CDC42BPA,CDK1,CEP 164,COL1A2,COL4A1,COX7A2,DFFB,DGKA,DLG1,DNAJA2,EDA2R ,EGFL6,EIF4G3,ENG,FAS,FDXR,FLRT2,FRMD4A,FYN,GNA13,GTF3 C2,HSPA1L,HTT,IFI16,IL4I1,INPP4A,KCNMA1,KIAA0368,LMO3,LP IN1,MAP2K4,MAP4,MAP4K1,MBNL2,MED13L,MIS18A,MLH1,M TP53 YH10,MYOF,NCAPG,NCAPH,NEDD8,NPEPPS,NR6A1,NUDT5,PAK 3,PCCA,PCDH7,PDCD6IP,PFN1,PIAS2,PIK3C3,PLAUR,PLK1,PPARD ,PPP4R2,PRKAG2,PRKAR2A,PRKCB,PRKCE,PRKD1,PRKG1,PSMD1, PSMD12,PTEN,PTK2,RAD50,RECQL4,RPS6KA2,SCP2,SMAD7,SRC ,STAT1,STIP1,TACC2,TCF7L2,THBS1,TIMP2,TJP1,TP63,TP73,TRIO ,TSC22D3,TULP4,UBE2C,UVRAG,XRCC5,ZNF365 Yap1 MYH11,MYOCD
FGF23 AKAP1,ATP1A2,SLC34A1 Factor
Growth TGFB3 ENG,ETS1,TJP1,ZEB2
miR‐182‐5p ADCY6,BCL2L14,CARD11,CDH4,COL11A2,NCAM1, NFASC,VWF miR‐185‐5p CDC42,RHOA miR‐30a‐3p ANKFY1,ARMC8,RSU1,SCAF11
miR‐92a‐3p CCND2,PTEN,TP63 mir‐145 FSCN1,MYH11,MYOCD
microRNA mir‐182 ADCY6,BCL2L14,CARD11,CDH4,COL11A2,NCAM1, NFASC,VWF mir‐183 PTEN,TAOK1 mir‐22 ARRB1,BMP7,MECOM mir‐8 E2F3,PTEN,ZEB2
118 SUPPLEMENT
Supplementary Table 9: Upstream Regulators associated to genes that show intragenic labeling of doxorubicin‐induced DSB
Molecule Upstream Target Molecules Type Regulator ATF1 CREB1,FTH1,INHBA,RHOA,TEAD1,THBS1 ATF2 CTNNB1,FN1,MAPK14,MYLK,PDGFRA,PPARG,PPARGC1A,PSEN1,PTEN,ZNF268 BACH1 CALM1 (includes others),EWSR1,FTH1,SQSTM1,TTC23 BMI1 DLC1,HK2,LOX,MYT1,NFKB1,NID1,P4HA1,SMAD3,VEZT ACTR3,ADK,ALG5,ATAD2,BIRC5,BRCA1,CALM1 (includes others),CASP9,CBX5,CDK1,CWC27,DUT,EYA3,HNRNPD,MAPK14,MCL1,MCM3,MF E2F1 AP1,MSH2,NCL,NCOA3,NDC80,PA2G4,PDK1,PIK3R1,PRIM2,PRKDC,PSAT1,PTPN4, RAD54L,RBBP4,RBFOX2,RBL1,REV3L,RRM1,RUVBL1,SAFB,SMARCA5,TOP2A,TRMT 13,TXNRD1,VRK1,WEE1,YWHAE ATRX,BARD1,BRCA1,CALM1 (includes others),CBX5,CDK1,CENPE,DUT,H2AFZ,LPAR1,MAP3K7,MCM3,MCM6,MSH2,NCL, E2F4 NDC80,NPM1,PKN2,PRIM2,PRKDC,PSAT1,RAD54L,RBBP4,RBL1,RRM1,SRRM2,TOP 2A,TRMT13,TTK ELK1 BMPR2,MCL1,MECOM,MYLK,PRKCA,RUNX2,THBS1,TIPARP ABCC4,ADD1,ARFGEF2,ARHGAP22,ARHGAP24,CDC42BPB,CDK5RAP2,CLIP1,CTNN B1,DAAM1,DIAPH2,DIAPH3,DIP2A,DOCK1,DOCK2,DSTN,ETS1,EXT1,MAGI1,MYO1 ERG D,MYO5A,NPHP1,NRG1,NUMB,PCNP,PPAP2B,PTPN11,PTPN4,RAB7A,RASA2,RPL2
2,RYR3,SLIT2,SVIL,TACC1,TMEM263,TSC22D3,UTRN,WDR91 FOXM1 ADAM17,BIRC5,CDK1,CDKN3,ESR1,LAMA4,LOX,SKP2,VCAN,ZEB1,ZEB2 ANGPT1,ANKRD17,APCDD1,CDX2,CTNNB1,DYRK1A,EFEMP1,GREB1,INHBA,NCKAP Regulator GMNN 1,NUDT21,PDGFRA,TEAD1,TGFB2,TWSG1,TXNRD1,WLS,YAP1 HAND2 ATP2A2,CCBE1,COL1A2,COL3A1,FAP,GLI3,GREM1,LOC102724428/SIK1,SHH CAP2,CNKSR2,DGKB,DLG2,DYRK1A,FGF13,GABRG3,GPHN,HIF1A,MAPK1,MAPK10, HDAC4 Tramscription NBEA,NTM,PRKCA,SYNGAP1 HMGA1 BRCA1,CAV2,DAB2,ELK1,INSIG1,KITLG,PPARG,STAT3 HR COL25A1,DYM,FGF13,IBTK,KCNIP1,MAGI2,TAF7L,UBR2 AP4S1,BRCA1,CDK1,DHCR24,DLGAP5,EPB41L1,INSIG1,ITGB1BP1,MAPK8IP3,MCM KDM5B 3,NDC80,PHF20,PSD3,PSIP1,PTPLA,SMC5,SNRPG,SPTSSA,TOP2A,TTK MDM2 FOXO3,HIF1A,IGF1R,KAT2B,MDM4,TOP2A MTDH KDM6A,PDCD11 ACACA,ADD1,ADK,AEBP2,AKAP12,ANGPT1,ANKRD17,APCDD1,ARF3,ARHGAP22,C AST,CCT3,CD44,CDK1,CDX2,CNTNAP2,COL15A1,COL1A2,COL3A1,COL4A1,COL4A2 ,COL5A1,COL5A2,COL6A3,COL8A1,CTNNB1,EFEMP1,EFTUD2,EIF2S1,EIF2S2,ENO1 ,F2R,FAP,FBN1,FECH,FN1,FSTL1,FTH1,GFPT1,GIGYF2,GLS,GLS2,GLUD1,GREB1,GT F2B,GTF2F2,HAPLN1,HERC5,HIF1A,HIVEP2,HK2,HSP90AA1,INHBA,IPO7,IQGAP2,IR MYC F6,ITGB1,JARID2,LOX,MAN2A1,MCM6,MFAP1,NCKAP1,NCL,NFAT5,NFATC3,NPM 1,NUDC,NUDT21,PDCD4,PDGFRA,PDK1,PGK1,PHF20,PHF21A,POLR2D,PREP,RAB1 0,RARS,RBBP4,RRM2B,RUVBL1,SCAMP1,SCEL,SDCBP,SKP2,ST3GAL1,ST3GAL4,TEA D1,TGFB2,THBS1,THBS2,TIMM10B,TNC,TNFRSF10B,TPP2,TWSG1,TXNRD1,WLS,Y AP1,YTHDC1 NEUROG1 ADD3,CEMIP,COL3A1,FAP,FN1,GREM1,INHBA,LGR5,PAPPA,SPOCK1,SULF1,THBS1, 119 SUPPLEMENT
TMTC2 ALG5,BIRC5,BRCA1,CASP9,CBX5,CDC25C,CDK1,CDX2,CWC27,GMPS,IGF1R,KPNA4, RB1 MAPK1,MFAP1,MTOR,PPARG,PPARGC1A,PRKAR1A,PTPN4,RBL1,SAFB,SMARCA5,Y WHAE,ZEB1,ZEB2 RBFOX2 PTBP2,PTEN SIAH2 ANGPT1,HIF1A,NCOR1 SMARCE1 BRCA1,BRCA2,CDK1,CENPE,SCN2A ANGPT1,ANKRD17,APCDD1,CDX2,CTNNB1,EFEMP1,GREB1,INHBA,NCKAP1,NUDT SOX1 21,PDGFRA,TEAD1,TGFB2,TWSG1,TXNRD1,WLS,YAP1 ADAM10,ADSS,CNOT6,EBF1,HILPDA,IL6ST,ITGB1,MSI2,MYO1B,NREP,RNF122,SET SOX11 MAR,STAT1 ANGPT1,ANKRD17,APCDD1,CDX2,CTNNB1,EFEMP1,GREB1,INHBA,NCKAP1,NUDT SOX3 21,PDGFRA,TEAD1,TGFB2,TWSG1,TXNRD1,WLS,YAP1 CDH11,COL4A1,COL4A2,COL5A1,COL5A2,COL6A3,DKK3,HIF1A,LAMC1,PRKCA,SM SPDEF AD3,TNC,ZEB1 TFAP4 CASP9,CD44,DYRK1A,LGR5 ABCB4,ACSL3,ADD3,AHCY,AKAP12,ARHGAP5,ASTN2,ASXL1,ATG4A,ATL3,ATXN1,B CAS3,BIRC5,BRCA1,C9,CAND1,CAP1,CDC16,CDC25C,CDC42BPA,CDC42EP3,CDK1, CDKN3,CHUK,COL1A2,COL3A1,COL4A1,COL5A2,COPB1,CSNK1D,CTNNB1,DHCR24 ,DICER1,DKK3,DLD,DLG1,DNAJA2,DNM1L,DSTN,DUT,EIF4G3,ENPP2,ESR1,ETFA,F2 R,FAM120A,FDXR,FGF2,FLRT2,FN1,FOXO3,FSTL1,GLS2,GNA13,GNA14,GSK3B,GSN ,GTF3C2,HADHA,HDLBP,HERC5,HIF1A,HK2,HSP90AA1,HSPG2,HTT,IDH1,IGF1R,IGF BP7,IL31RA,IL7,INHBA,INPP4A,IPO7,IPO9,KAT2B,KCNMA1,KIAA0368,KIF23,KITLG, KPNB1,LATS2,LIMA1,LPP,MAN2A1,MAP2K4,MAP4,MAPK1,MCL1,MCM3,MCM6, TP53 MDM4,MED13L,MICALL1,MPI,MSH2,MTA1,MTDH,MYBL1,MYH9,MYO6,NDC80,N LRC4,NR6A1,NRAP,NUP153,P4HA1,PAK3,PANK1,PARD6B,PARK7,PCCA,PDCD6IP,P DE4B,PDGFRA,PDK1,PIAS2,PIK3C3,PIK3R4,PLOD2,POLE2,PPM1D,PPP1R13L,PRKAR 2A,PRKCZ,PRKG1,PSMD1,PSMD12,PTEN,PTK2,PTPN11,PTPRM,PVRL3,RAD23A,RA D50,RAF1,RBBP4,RBL1,RFC1,RGS12,ROBO1,RPN1,RPN2,RPS6KB1,RRM1,RRM2B,R UNX2,SCMH1,SEC62,SHROOM3,SIRT1,SMAD6,SMC3,SMURF2,SON,SPATA18,SQLE ,SRSF3,STAT1,STAU1,STK17A,TANK,TDG,TGFB2,THBS1,TNFRSF10B,TOP2A,TPX2,T RIO,TSC22D3,TTK,UBE2B,UBL3,UIMC1,ULK2,USO1,USP14,UVRAG,VCAN,VCL,VRK1 ,XRCC5,ZEB1,ZFP36L1 ATG4A,BRCA1,CAST,CDC25C,CDC42,CDK1,COL4A1,COL5A1,CYR61,DICER1,F2R,FB TP63 N1,FN1,FOXO3,HMGA2,IGFBP7,KIF23,PNPT1,PRKCZ,PTEN,SPATA18,THBS1,TNC,U LK2,UVRAG,WEE1,ZEB1 TWIST2 CD44,CTNNB1,FN1,ZEB1,ZEB2 ARCN1,ATP2A2,COPB1,COPB2,DERL1,DNAJC10,EDEM1,ERLEC1,ESR1,ETS1,FN1,G XBP1 OLGA3,GOLGA4,GORASP2,HM13,KDELR2,PDIA3,PPIB,RPN1,RPN2,SDF2L1,SEC22B ,SEC23A,SEC31A,SEC61G,SEC63,SMC3,SRP54,SRP68,SSR1,TXN,USO1,XRCC6 ADD3,AKTIP,CD44,ERRFI1,ETV5,FRMD4B,INHBA,SCEL,SLC39A10,SMARCA1,TXNRD
BMP6 1,VCL,ZEB1 CD44,COL4A2,COL8A1,CTNNB1,ENO1,ESR1,FN1,HAPLN1,HIF1A,KITLG,LIMS1,P4H Factor Growth CTGF A1,PDE7A,PI4K2B,TMOD3,WNK1 120 SUPPLEMENT
BIRC5,CAPZA2,CDH11,CDX2,CHRM3,COL1A2,CTNNB1,ENO1,FGF2,FN1,FTH1,NF1, FGF2 PGK1,PPARG,RUNX2,SQSTM1,ST3GAL1,ST3GAL4,TGFB2,TOP2A WISP2 CD44,ESR1,FN1,HIF1A,IGFBP7,SMAD3,TGFBR1,ZEB1 miR‐128‐3p TGFBR1,WEE1 miR‐185‐5p CDC42,RHOA miR‐192‐5p BIRC5,IGF1R,ZEB1 miR‐199a‐5p BIRC5,DYRK1A,ETS1 miR‐19b‐3p BIRC5,PTEN miR‐21‐5p MSH2,PDCD4,PIK3R1,PTEN,STAT3 miR‐217‐5p ACACA,PPARGC1A,SIRT1,TRPS1
miR‐29b‐3p COL1A2,COL5A2,DCP2,HDAC4,LAMC1,MAPRE2,NAV3,RERE,ZFP36L1 miR‐30a‐3p ARMC8,KIF1B,LIMA1,MCCC2,RAB8B,SCAF11,WDR44
microRNA mir‐1 ADD1,ANXA2,CAP1,FN1,PICALM,RASA1,RHEB,XPO6 mir‐144 KLF7,RAB14,RAC1,RASA1,TLK2,YWHAZ mir‐154 BIRC5,PTEN mir‐181 MCL1,PTEN,SPTLC1 mir‐199 DYRK1A,EGLN1,HIF1A,SIRT1 mir‐22 BMPR1B,EDC3,MECOM,YWHAZ mir‐451 KLF7,RAB14,RAC1,RASA1,TLK2,YWHAZ mir‐8 ITGB1,PTEN,TGFB2,YAP1,ZEB1,ZEB2,ZFPM2
Supplementary Table 10: Distribution of DSB sites identified in the three passages. Chr: Chromosome, TSS: Transcription Start Site
DSB Frequency
in passage
of
Exon Gene
[bp]
4 6 8 Change Gene Chr
of Position TSS Radiation Treament Fold Intron/ Upstream RefSeq Downstream
+IR Ctr 0.009 0.002 0.001 0.11 6 124852660 0 In3 0 NKAIN2 +IR Ctr 0.037 0.014 0.005 0.14 8 71528210 7606 0 TRAM1 +IR Ctr 0.028 0.024 0.013 0.46 10 19146063 0 179123 ARL5B +IR Ctr 0.019 0.012 0.013 0.68 15 39600267 0 53219 C15orf54 +IR Ctr 0.009 0.035 0.007 0.78 6 160280122 0 38387 PNLDC1 +IR Ctr 0.028 0.002 0.056 2.00 2 237175619 2631 0 ASB18 +IR M 0.065 0.002 0.003 0.05 7 39788845 0 41130 RALA +IR M 0.028 0.889 0.002 0.07 2 236001953 0 37597 SH3BP4 +IR M 0.009 0.011 0.001 0.11 1 200996263 3435 0 KIF21B +IR M 0.009 0.002 0.004 0.44 14 29163728 72559 0 FOXG1 +IR M 0.009 0.002 0.005 0.56 20 25383405 4918 0 GINS1 +IR M 0.009 0.006 0.008 0.89 12 69593947 39370 0 CPSF6 +IR M 0.046 0.023 0.057 1.24 6 87451391 195633 0 HTR1E +IR M 0.019 0.043 0.411 21.63 X 40608992 14188 0 MED14 +IR N 0.009 0.002 0.001 0.11 19 22248131 0 In1 0 ZNF257 +IR N 0.009 0.002 0.001 0.11 17 60848501 0 In3 0 MARCH10 121 SUPPLEMENT
+IR N 0.009 0.002 0.001 0.11 6 100799124 0 37627 SIM1 +IR N 0.009 0.002 0.001 0.11 3 194496759 0 86995 FAM43A +IR N 0.019 0.011 0.011 0.58 2 220242059 0 In13 0 DNPEP +IR N 0.009 0.049 0.009 1.00 2 9428234 0 In2 0 ASAP2 +IR N 0.019 0.014 0.039 2.05 6 75699157 0 94886 COL12A1 +IR N 0.009 0.040 0.021 2.33 8 63766653 0 In4 0 NKAIN3 +IR N 0.009 0.163 0.032 3.56 X 142717350 0 Ex2 0 SLITRK4 +IR N 0.009 0.002 0.034 3.78 5 132801252 0 In3 0 FSTL4 +IR N 0.019 0.011 0.340 17.89 12 26980249 0 In1 0 ITPR2 +IR MN 0.046 0.002 0.001 0.02 16 24990921 0 In1 0 ARHGAP17 +IR MN 0.093 0.041 0.004 0.04 9 84828995 151991 0 FAM75B +IR MN 0.083 0.005 0.005 0.06 16 79615445 0 12301 MAF +IR MN 0.009 0.002 0.001 0.11 7 116876671 0 6598 ST7 +IR MN 0.009 0.002 0.001 0.11 2 138258041 0 In14 0 THSD7B +IR MN 0.009 0.006 0.001 0.11 5 151230348 0 In7 0 GLRA1 +IR MN 0.037 0.018 0.005 0.14 20 25686101 8632 0 ZNF337 ‐IR Ctr 0.120 0.032 0.001 0.01 13 59624112 0 615613 DIAPH3 ‐IR Ctr 0.009 0.002 0.004 0.44 8 103799355 0 39182 AZIN1 ‐IR M 0.009 0.006 0.001 0.11 3 155453796 0 26605 C3orf33 ‐IR N 0.019 0.002 0.001 0.05 16 71576571 0 4158 CHST4 ‐IR N 0.009 0.002 0.001 0.11 1 12156057 0 In2 0 TNFRSF8 ‐IR N 0.009 0.002 0.001 0.11 21 18509311 376019 0 CXADR ‐IR N 0.009 0.002 0.001 0.11 14 68699623 0 In7 0 RAD51L1 ‐IR N 0.009 0.002 0.004 0.44 2 137507818 240644 0 THSD7B ‐IR MN 0.037 0.002 0.001 0.03 12 49271656 12003 0 RND1 ‐IR MN 0.009 0.002 0.001 0.11 2 40663201 0 In1 0 SLC8A1 ‐IR MN 0.009 0.002 0.001 0.11 5 96442660 0 In3 0 LIX1
Supplementary Table 11: Number and position of the top ten DSB sites over time. TSS: Transcription Start Site; TTS: Transcription Termination Site; +IR: irradiated, Ctr: cells with active NHEJ‐repair activity; MN: Mirin + Nu7441 treatment; TSS: Transcription Start Site; TTS: Transcription Termination Site Upstream of TSS In Gene Downstream of TTS Radiation Treatment p4 p6 p8 p4 p6 p8 p4 p6 p8 +IR Ctr 4 3 3 3 4 5 3 3 2 +IR +Mirin 5 3 5 3 6 2 2 1 3 +IR +Nu7441 6 1 2 2 5 4 2 4 4 +IR +MN 2 2 3 5 6 3 3 3 4
122 SUPPLEMENT
Supplementary Table 12: Association of radiation‐related DSB hotspots with neoplasia and cancer
Neo‐ Symbol Entrez Gene Name Type(s) plasia Cancer AAMDC adipogenesis associated, Mth938 domain containing other ARMC1 armadillo repeat containing 1 other X ASB1 ankyrin repeat and SOCS box containing 1 transcription regulator ATP6V0A1 ATPase, H+ transporting, lysosomal V0 subunit a1 transporter X X ATP6V1H ATPase, H+ transporting, lysosomal 50/57kDa, V1 transporter subunit H BCCIP BRCA2 and CDKN1A interacting protein other X X BCL11B B‐cell CLL/lymphoma 11B (zinc finger protein) other X X C11orf80 chromosome 11 open reading frame 80 other X X CARD11 caspase recruitment domain family, member 11 kinase X X CD93 CD93 molecule other X X CDH10 cadherin 10, type 2 (T2‐cadherin) other X X CEBPZ CCAAT/enhancer binding protein (C/EBP), zeta other X X CLMN calmin (calponin‐like, transmembrane) other X X CNOT1 CCR4‐NOT transcription complex, subunit 1 other X X CNTNAP2 contactin associated protein‐like 2 other X X CTNNA2 catenin (cadherin‐associated protein), alpha 2 other X X CWH43 cell wall biogenesis 43 C‐terminal homolog (S. other X X cerevisiae) CYP39A1 cytochrome P450, family 39, subfamily A, polypeptide 1 enzyme EDIL3 EGF‐like repeats and discoidin I‐like domains 3 other X X EDN1 endothelin 1 cytokine X X ELF4 E74‐like factor 4 (ets domain transcription factor) transcription regulator EMR2 egf‐like module containing, mucin‐like, hormone other X X receptor‐like 2 ERI1 exoribonuclease 1 enzyme X X ERICH3 glutamate‐rich 3 other X X FAM120C family with sequence similarity 120C other X X FAM174A family with sequence similarity 174, member A other X X FCGR1B Fc fragment of IgG, high affinity Ib, receptor (CD64) transmem‐brane X receptor FPGT‐ FPGT‐TNNI3K readthrough other X TNNI3K FRMPD4 FERM and PDZ domain containing 4 other X X GALNT5 polypeptide N‐acetylgalactosaminyltransferase 5 enzyme X X GINS3 GINS complex subunit 3 (Psf3 homolog) other X X H2AFY H2A histone family, member Y other X X HIST1H1C histone cluster 1, H1c other X X HIST1H2AC histone cluster 1, H2ac other 123 SUPPLEMENT
HIST1H3B histone cluster 1, H3b other X X HIVEP3 human immunodeficiency virus type I enhancer binding transcription X X protein 3 regulator HS6ST1 heparan sulfate 6‐O‐sulfotransferase 1 enzyme X X HTRA2 HtrA serine peptidase 2 peptidase X X IL17B interleukin 17B cytokine X INTS4 integrator complex subunit 4 other X X IRS2 insulin receptor substrate 2 enzyme X X KCTD20 potassium channel tetramerization domain containing ion channel X X 20 KIAA2013 KIAA2013 other X LACE1 lactation elevated 1 other X MIIP migration and invasion inhibitory protein other MORN1 MORN repeat containing 1 other X X MYT1L myelin transcription factor 1‐like transcription X X regulator NDRG4 NDRG family member 4 other X X NKX2‐3 NK2 homeobox 3 transcription regulator NLGN1 neuroligin 1 enzyme X X OR5AN1 olfactory receptor, family 5, subfamily AN, member 1 G‐protein X X coupled receptor OR7A17 olfactory receptor, family 7, subfamily A, member 17 G‐protein X X coupled receptor OR7A5 olfactory receptor, family 7, subfamily A, member 5 G‐protein X X coupled receptor PDE4D phosphodiesterase 4D, cAMP‐specific enzyme X X PEX10 peroxisomal biogenesis factor 10 other X PHF8 PHD finger protein 8 enzyme X X PLA2G7 phospholipase A2, group VII (platelet‐activating factor enzyme X X acetylhydrolase, plasma) PLAUR plasminogen activator, urokinase receptor transmem‐brane X X receptor PNLDC1 poly(A)‐specific ribonuclease (PARN)‐like domain other X X containing 1 PPP1R3B protein phosphatase 1, regulatory subunit 3B other X PRKX protein kinase, X‐linked kinase X X RPIA ribose 5‐phosphate isomerase A enzyme X X RSF1 remodeling and spacing factor 1 transcription X X regulator SCLY selenocysteine lyase enzyme X X SETBP1 SET binding protein 1 other X X SIPA1L2 signal‐induced proliferation‐associated 1 like 2 other X X SPTBN2 spectrin, beta, non‐erythrocytic 2 other X X 124 SUPPLEMENT
STAT3 signal transducer and activator of transcription 3 (acute‐ transcription X X phase response factor) regulator STAT5B signal transducer and activator of transcription 5B transcription X X regulator STK38 serine/threonine kinase 38 kinase SULT6B1 sulfotransferase family, cytosolic, 6B, member 1 other X X SYNE3 spectrin repeat containing, nuclear envelope family other X X member 3 TAS2R41 taste receptor, type 2, member 41 G‐protein X coupled receptor TCP1 t‐complex 1 other X X TEX36 testis expressed 36 other X TFPI tissue factor pathway inhibitor (lipoprotein‐associated other X X coagulation inhibitor) TIFAB TRAF‐interacting protein with forkhead‐associated other X X domain, family member B TNFRSF8 tumor necrosis factor receptor superfamily, member 8 transmem‐brane X X receptor TOP1 topoisomerase (DNA) I enzyme X X TRIM38 tripartite motif containing 38 other X X UBE2E2 ubiquitin‐conjugating enzyme E2E 2 enzyme UBE2F ubiquitin‐conjugating enzyme E2F (putative) enzyme WFDC2 WAP four‐disulfide core domain 2 other X X XRCC1 X‐ray repair complementing defective repair in Chinese other X X hamster cells 1 ZMAT2 zinc finger, matrin‐type 2 other X X ZNF280C zinc finger protein 280C other X X ZNF33B zinc finger protein 33B transcription X X regulator ZYX zyxin other X X
Supplementary Table 13: Association of doxorubicin‐related DSB hotspots with neoplasia and cancer
Symbol Entrez Gene Name Type(s) Neo‐ Cancer plasia AACS acetoacetyl‐CoA synthetase enzyme X X AAMP angio‐associated, migratory cell protein other X X AARS alanyl‐tRNA synthetase enzyme X X ABCA12 ATP‐binding cassette, sub‐family A (ABC1), member 12 transporter X X ADPRHL1 ADP‐ribosylhydrolase like 1 enzyme X X AGBL2 ATP/GTP binding protein‐like 2 enzyme X X AKAP12 A kinase (PRKA) anchor protein 12 transporter X X ANKFY1 ankyrin repeat and FYVE domain containing 1 transcription X X regulator ANKRD12 ankyrin repeat domain 12 other X X 125 SUPPLEMENT
AP4E1 adaptor‐related protein complex 4, epsilon 1 subunit other X X ARHGEF11 Rho guanine nucleotide exchange factor (GEF) 11 other X X ARID4A AT rich interactive domain 4A (RBP1‐like) transcription X X regulator ARIH1 ariadne RBR E3 ubiquitin protein ligase 1 enzyme X X ARL4D ADP‐ribosylation factor‐like 4D enzyme X X ARMC9 armadillo repeat containing 9 other X X ASB3/GPR7 ankyrin repeat and SOCS box containing 3 transcription X 5‐ASB3 regulator ASTN2 astrotactin 2 other X X ASZ1 ankyrin repeat, SAM and basic leucine zipper domain transcription X X containing 1 regulator ATP10D ATPase, class V, type 10D transporter X X ATP13A3 ATPase type 13A3 transporter X X ATP13A4 ATPase type 13A4 transporter X B4GALNT2 beta‐1,4‐N‐acetyl‐galactosaminyl transferase 2 enzyme X BAZ1B bromodomain adjacent to zinc finger domain, 1B transcription X X regulator BIRC6 baculoviral IAP repeat containing 6 enzyme X X BRWD1 bromodomain and WD repeat domain containing 1 transcription X X regulator C14orf166 chromosome 14 open reading frame 166 other X X C1QTNF4 C1q and tumor necrosis factor related protein 4 other X C5orf34 chromosome 5 open reading frame 34 other X X C6orf211 chromosome 6 open reading frame 211 other X X CA12 carbonic anhydrase XII enzyme X X CALD1 caldesmon 1 other X X CANX calnexin other X X CASP9 caspase 9, apoptosis‐related cysteine peptidase peptidase X X CCAR1 cell division cycle and apoptosis regulator 1 transcription X X regulator CCBE1 collagen and calcium binding EGF domains 1 other X CCDC12 coiled‐coil domain containing 12 other X X CCNT1 cyclin T1 transcription X X regulator CDAN1 codanin 1 other X CDC25C cell division cycle 25C phosphatase X CDC27 cell division cycle 27 other X X CDK12 cyclin‐dependent kinase 12 kinase X X CDK13 cyclin‐dependent kinase 13 kinase X X CELF1 CUGBP, Elav‐like family member 1 translation X X regulator CEP97 centrosomal protein 97kDa other X CFAP36 cilia and flagella associated protein 36 other X X 126 SUPPLEMENT
CHD8 chromodomain helicase DNA binding protein 8 enzyme X X CHRDL2 chordin‐like 2 other X CHUK conserved helix‐loop‐helix ubiquitous kinase kinase X X CLNS1A chloride channel, nucleotide‐sensitive, 1A ion channel X CLP1 cleavage and polyadenylation factor I subunit 1 other X X CMSS1 cms1 ribosomal small subunit homolog (yeast) other X CNIH4 cornichon family AMPA receptor auxiliary protein 4 other X X CNTRL centriolin transcription X X regulator COL3A1 collagen, type III, alpha 1 other X X COL5A2 collagen, type V, alpha 2 other X X COL8A1 collagen, type VIII, alpha 1 other X CORIN corin, serine peptidase peptidase X X CPNE3 copine III kinase X X CPQ carboxypeptidase Q peptidase X X CSNK1G1 casein kinase 1, gamma 1 kinase X X CXCR1 chemokine (C‐X‐C motif) receptor 1 G‐protein X coupled receptor CXCR2 chemokine (C‐X‐C motif) receptor 2 G‐protein X X coupled receptor DAGLB diacylglycerol lipase, beta enzyme X DAPP1 dual adaptor of phosphotyrosine and 3‐ other X phosphoinositides DCUN1D4 DCN1, defective in cullin neddylation 1, domain other X containing 4 DDX50 DEAD (Asp‐Glu‐Ala‐Asp) box polypeptide 50 enzyme X X DENND5A DENN/MADD domain containing 5A other X X DIP2A DIP2 disco‐interacting protein 2 homolog A (Drosophila) transcription X X regulator DIS3 DIS3 exosome endoribonuclease and 3’‐5’ enzyme X X exoribonuclease DNA2 DNA replication helicase/nuclease 2 enzyme X X DNAAF2 dynein, axonemal, assembly factor 2 other X X DNAJA3 DnaJ (Hsp40) homolog, subfamily A, member 3 other X DNAJB6 DnaJ (Hsp40) homolog, subfamily B, member 6 transcription X X regulator DNAJC10 DnaJ (Hsp40) homolog, subfamily C, member 10 enzyme X X DNAJC16 DnaJ (Hsp40) homolog, subfamily C, member 16 other X DOCK11 dedicator of cytokinesis 11 other X X DST dystonin other X X DTNB dystrobrevin, beta other X X DYNC2H1 dynein, cytoplasmic 2, heavy chain 1 other X X 127 SUPPLEMENT
DYNLL1 dynein, light chain, LC8‐type 1 other X X ECD ecdysoneless homolog (Drosophila) transcription X X regulator EFCAB13 EF‐hand calcium binding domain 13 other X X EGR1 early growth response 1 transcription X X regulator EHMT1 euchromatic histone‐lysine N‐methyltransferase 1 transcription X X regulator EIF1 eukaryotic translation initiation factor 1 translation X regulator EIF2AK2 eukaryotic translation initiation factor 2‐alpha kinase 2 kinase X X EIF2S1 eukaryotic translation initiation factor 2, subunit 1 alpha, translation X X 35kDa regulator ENPP2 ectonucleotide pyrophosphatase/phosphodiesterase 2 enzyme X X ERICH5 glutamate‐rich 5 other X X ERLEC1 endoplasmic reticulum lectin 1 other X X ETV3L ets variant 3‐like other X X EXT1 exostosin glycosyltransferase 1 enzyme X X F2R coagulation factor II (thrombin) receptor G‐protein X X coupled receptor F2RL1 coagulation factor II (thrombin) receptor‐like 1 G‐protein X X coupled receptor FAM149B1 family with sequence similarity 149, member B1 other X X FAM179B family with sequence similarity 179, member B other X X FAM186B family with sequence similarity 186, member B other X FAM71D family with sequence similarity 71, member D other X FAM98A family with sequence similarity 98, member A other X X FBXO22 F‐box protein 22 enzyme X X FILIP1 filamin A interacting protein 1 other X X FMO4 flavin containing monooxygenase 4 enzyme X FN1 fibronectin 1 enzyme X X FNBP4 formin binding protein 4 other X X FRMD6 FERM domain containing 6 other X X FYTTD1 forty‐two‐three domain containing 1 other X X GABBR1 gamma‐aminobutyric acid (GABA) B receptor, 1 G‐protein X X coupled receptor GAST gastrin other X X GCN1L1 GCN1 general control of amino‐acid synthesis 1‐like 1 translation X X (yeast) regulator GEMIN5 gem (nuclear organelle) associated protein 5 other X X GIP gastric inhibitory polypeptide other X 128 SUPPLEMENT
GLS glutaminase enzyme X X GNPNAT1 glucosamine‐phosphate N‐acetyltransferase 1 enzyme X X GORASP2 golgi reassembly stacking protein 2, 55kDa other X X GP5 glycoprotein V (platelet) other X X GSN gelsolin other X X HEATR5B HEAT repeat containing 5B other X X HMCN1 hemicentin 1 other X X HMGA2 high mobility group AT‐hook 2 enzyme X X HMGXB4 HMG box domain containing 4 other X X HNRNPH3 heterogeneous nuclear ribonucleoprotein H3 (2H9) other X X IGF2BP1 insulin‐like growth factor 2 mRNA binding protein 1 translation X X regulator IGF2BP2 insulin‐like growth factor 2 mRNA binding protein 2 translation X X regulator IGFBP7 insulin‐like growth factor binding protein 7 transporter X X IMPG1 interphotoreceptor matrix proteoglycan 1 other X X IPO7 importin 7 transporter X X IQGAP2 IQ motif containing GTPase activating protein 2 other X X KANSL1 KAT8 regulatory NSL complex subunit 1 other X X KANSL2 KAT8 regulatory NSL complex subunit 2 other X X KCTD20 potassium channel tetramerization domain containing 20 kinase X X KDELR2 KDEL (Lys‐Asp‐Glu‐Leu) endoplasmic reticulum protein other X X retention receptor 2 KDM3B lysine (K)‐specific demethylase 3B other X X KIAA0196 KIAA0196 other X KIAA0226 KIAA0226 other X X KIAA0586 KIAA0586 other X KIAA2022 KIAA2022 enzyme X X KIF13A kinesin family member 13A transporter X X KLF14 Kruppel‐like factor 14 other X X KLHDC2 kelch domain containing 2 other X X KLHL35 kelch‐like family member 35 other X X KMT2A lysine (K)‐specific methyltransferase 2A transcription X X regulator KPNA4 karyopherin alpha 4 (importin alpha 3) transporter X X LMLN leishmanolysin‐like (metallopeptidase M8 family) peptidase X X LOX lysyl oxidase enzyme X X LRCH3 leucine‐rich repeats and calponin homology (CH) domain other X X containing 3 LRP5 low density lipoprotein receptor‐related protein 5 transmem‐ X X brane receptor LRR1 leucine rich repeat protein 1 enzyme X X LRRC57 leucine rich repeat containing 57 other X X 129 SUPPLEMENT
LTN1 listerin E3 ubiquitin protein ligase 1 enzyme X X MAML1 mastermind‐like 1 (Drosophila) transcription X X regulator MAP3K2 mitogen‐activated protein kinase kinase kinase 2 kinase X X MAP4K5 mitogen‐activated protein kinase kinase kinase kinase 5 kinase X X MBTD1 mbt domain containing 1 other X X MCM3AP minichromosome maintenance complex component 3 other X X associated protein MDM4 MDM4, p53 regulator other X X METTL16 methyltransferase like 16 other X X MFAP1 microfibrillar‐associated protein 1 other X X MGRN1 mahogunin ring finger 1, E3 ubiquitin protein ligase enzyme X X MKLN1 muskelin 1, intracellular mediator containing kelch motifs other X X MLEC malectin other X X MPP5 membrane protein, palmitoylated 5 (MAGUK p55 kinase X X subfamily member 5) MRPL22 mitochondrial ribosomal protein L22 other X X MRPS16 mitochondrial ribosomal protein S16 other MSRB3 methionine sulfoxide reductase B3 other X X MTCH2 mitochondrial carrier 2 other X MYO6 myosin VI other X X MYO9A myosin IXA enzyme X X MZT1 mitotic spindle organizing protein 1 other X X NBR1 neighbor of BRCA1 gene 1 other X X NCBP2 nuclear cap binding protein subunit 2, 20kDa other X NCOA6 nuclear receptor coactivator 6 transcription X X regulator NCOR1 nuclear receptor corepressor 1 transcription X X regulator NDUFV2 NADH dehydrogenase (ubiquinone) flavoprotein 2, 24kDa enzyme X X NEMF nuclear export mediator factor other X NFAT5 nuclear factor of activated T‐cells 5, tonicity‐responsive transcription X X regulator NLK nemo‐like kinase kinase X X NLRC4 NLR family, CARD domain containing 4 other X X NRD1 nardilysin (N‐arginine dibasic convertase) peptidase X X NRG4 neuregulin 4 growth factor NSMCE2 non‐SMC element 2, MMS21 homolog (S. cerevisiae) enzyme X X NUP153 nucleoporin 153kDa transporter X X P4HA1 prolyl 4‐hydroxylase, alpha polypeptide I enzyme X X PAFAH1B1 platelet‐activating factor acetylhydrolase 1b, regulatory enzyme X subunit 1 (45kDa) PAIP1 poly(A) binding protein interacting protein 1 translation X X regulator 130 SUPPLEMENT
PAK2 p21 protein (Cdc42/Rac)‐activated kinase 2 kinase X X PANK3 pantothenate kinase 3 kinase X X PAPPA pregnancy‐associated plasma protein A, pappalysin 1 peptidase X X PCDH10 protocadherin 10 other X X PDIA3 protein disulfide isomerase family A, member 3 peptidase X X PDPR pyruvate dehydrogenase phosphatase regulatory subunit enzyme X X PIBF1 progesterone immunomodulatory binding factor 1 other X X PIK3C2B phosphatidylinositol‐4‐phosphate 3‐kinase, catalytic kinase X X subunit type 2 beta PIK3C3 phosphatidylinositol 3‐kinase, catalytic subunit type 3 kinase X X PKD2L1 polycystic kidney disease 2‐like 1 ion channel X PLEK2 pleckstrin 2 other X X PLEKHM2 pleckstrin homology domain containing, family M (with other X X RUN domain) member 2 PNKD paroxysmal nonkinesigenic dyskinesia other X X PNPT1 polyribonucleotide nucleotidyltransferase 1 enzyme X POLD3 polymerase (DNA‐directed), delta 3, accessory subunit transcription X X regulator POLE2 polymerase (DNA directed), epsilon 2, accessory subunit enzyme X X POLR2B polymerase (RNA) II (DNA directed) polypeptide B, enzyme X X 140kDa POP1 processing of precursor 1, ribonuclease P/MRP subunit (S. enzyme X X cerevisiae) PPP1CB protein phosphatase 1, catalytic subunit, beta isozyme phosphatase X X PPP1R15B protein phosphatase 1, regulatory subunit 15B phosphatase X X PRG4 proteoglycan 4 other X X PRKAR1A protein kinase, cAMP‐dependent, regulatory, type I, alpha kinase X X PRRC2C proline‐rich coiled‐coil 2C other X PSEN1 presenilin 1 peptidase X X PSMD1 proteasome (prosome, macropain) 26S subunit, non‐ other X X ATPase, 1 PSMD14 proteasome (prosome, macropain) 26S subunit, non‐ peptidase X X ATPase, 14 PVRL3 poliovirus receptor‐related 3 other X X RAB14 RAB14, member RAS 130ncogenes family enzyme X X RAC1 ras‐related C3 botulinum toxin substrate 1 (rho family, enzyme X X small GTP binding protein Rac1) RANBP9 RAN binding protein 9 other X X RARS arginyl‐tRNA synthetase enzyme X X RASGRP3 RAS guanyl releasing protein 3 (calcium and DAG‐ other X X regulated) RBM25 RNA binding motif protein 25 other X X RBM5 RNA binding motif protein 5 other X X RBM6 RNA binding motif protein 6 other X X 131 SUPPLEMENT
REST RE1‐silencing transcription factor transcription X X regulator RFC1 replication factor C (activator 1) 1, 145kDa transcription X X regulator RHBDL3 rhomboid, veinlet‐like 3 (Drosophila) peptidase X X RLIM ring finger protein, LIM domain interacting enzyme X X RMDN1 regulator of microtubule dynamics 1 other X X RMND1 required for meiotic nuclear division 1 homolog (S. other X X cerevisiae) RNF10 ring finger protein 10 enzyme X X RNF139 ring finger protein 139 enzyme X X RNF169 ring finger protein 169 other X RNF216 ring finger protein 216 other X X RPGRIP1 retinitis pigmentosa GTPase regulator interacting protein other X X 1 RPLP0 ribosomal protein, large, P0 other X X RPS3 ribosomal protein S3 enzyme X RSF1 remodeling and spacing factor 1 transcription X X regulator RSPRY1 ring finger and SPRY domain containing 1 other X X RUFY2 RUN and FYVE domain containing 2 other X X SCD5 stearoyl‐CoA desaturase 5 enzyme X X SEC23A Sec23 homolog A (S. cerevisiae) transporter X X SEC31A SEC31 homolog A (S. cerevisiae) other X X SENP2 SUMO1/sentrin/SMT3 specific peptidase 2 peptidase X X SENP5 SUMO1/sentrin specific peptidase 5 peptidase X X SENP6 SUMO1/sentrin specific peptidase 6 peptidase X X SERF2 small EDRK‐rich factor 2 other X X SERPINH1 serpin peptidase inhibitor, clade H (heat shock protein other X X 47), member 1, (collagen binding protein 1) SETD2 SET domain containing 2 enzyme X X SGCB sarcoglycan, beta (43kDa dystrophin‐associated other X X glycoprotein) SIRT4 sirtuin 4 enzyme X X SIRT5 sirtuin 5 enzyme X SLC25A17 solute carrier family 25 (mitochondrial carrier; transporter X X peroxisomal membrane protein, 34kDa), member 17 SMEK2 SMEK homolog 2, suppressor of mek1 (Dictyostelium) other X X SMG6 SMG6 nonsense mediated mRNA decay factor enzyme X X SNX1 sorting nexin 1 transporter X SPAG9 sperm associated antigen 9 other X X SPATA18 spermatogenesis associated 18 other X X SPATS2 spermatogenesis associated, serine‐rich 2 other X X SPDYA speedy/RINGO cell cycle regulator family member A other X X 132 SUPPLEMENT
SQLE squalene epoxidase enzyme X X SQSTM1 sequestosome 1 transcription X regulator SRFBP1 serum response factor binding protein 1 other X X SRSF3 serine/arginine‐rich splicing factor 3 other X ST7 suppression of tumorigenicity 7 other X STAT1 signal transducer and activator of transcription 1, 91kDa transcription X X regulator STAT4 signal transducer and activator of transcription 4 transcription X X regulator STK3 serine/threonine kinase 3 kinase X X STOX1 storkhead box 1 other X STRN striatin, calmodulin binding protein other X SUGCT succinyl‐CoA:glutarate‐CoA transferase enzyme X X SUV420H1 suppressor of variegation 4‐20 homolog 1 (Drosophila) enzyme X X SYNE1 spectrin repeat containing, nuclear envelope 1 other X X TACC1 transforming, acidic coiled‐coil containing protein 1 other X X TAF2 TAF2 RNA polymerase II, TATA box binding protein (TBP)‐ transcription X X associated factor, 150kDa regulator TATDN1 TatD Dnase domain containing 1 other X X TBC1D23 TBC1 domain family, member 23 other X X TBK1 TANK‐binding kinase 1 kinase X X TGM5 transglutaminase 5 enzyme X X TGM7 transglutaminase 7 enzyme X THRB thyroid hormone receptor, beta ligand‐ X X dependent nuclear receptor TLK1 tousled‐like kinase 1 kinase X X TMCO3 transmembrane and coiled‐coil domains 3 transporter X X TMEM44 transmembrane protein 44 other X X TNPO1 transportin 1 transporter X X TOMM70A translocase of outer mitochondrial membrane 70 transporter X X homolog A (S. cerevisiae) TP53BP1 tumor protein p53 binding protein 1 transcription X X regulator TPR translocated promoter region, nuclear basket protein transporter X X TRIM37 tripartite motif containing 37 enzyme X X TRMT61B tRNA methyltransferase 61 homolog B (S. cerevisiae) enzyme X TTC19 tetratricopeptide repeat domain 19 other X UBE2Q2 ubiquitin‐conjugating enzyme E2Q family member 2 enzyme X X UBE3C ubiquitin protein ligase E3C enzyme X X UBE4A ubiquitination factor E4A enzyme X X UBR2 ubiquitin protein ligase E3 component n‐recognin 2 enzyme X X 133 SUPPLEMENT
USP15 ubiquitin specific peptidase 15 peptidase X X USP16 ubiquitin specific peptidase 16 peptidase X X USP3 ubiquitin specific peptidase 3 peptidase X UTP18 UTP18 small subunit (SSU) processome component other X homolog (yeast) VIT vitrin other X X VPS13C vacuolar protein sorting 13 homolog C (S. cerevisiae) other X X VPS13D vacuolar protein sorting 13 homolog D (S. cerevisiae) other X X WDR19 WD repeat domain 19 other X X WDR76 WD repeat domain 76 other X X WWC1 WW and C2 domain containing 1 transcription X X regulator WWP1 WW domain containing E3 ubiquitin protein ligase 1 enzyme X X WWP2 WW domain containing E3 ubiquitin protein ligase 2 enzyme X X XPNPEP3 X‐prolyl aminopeptidase (aminopeptidase P) 3, putative peptidase X X XPOT exportin, tRNA other X X XRCC5 X‐ray repair complementing defective repair in Chinese enzyme X X hamster cells 5 (double‐strand‐break rejoining) YBEY ybeY metallopeptidase (putative) other X X YIPF4 Yip1 domain family, member 4 other X X YTHDF2 YTH domain family, member 2 other X ZBTB11 zinc finger and BTB domain containing 11 other X X ZBTB2 zinc finger and BTB domain containing 2 other X X ZDHHC5 zinc finger, DHHC‐type containing 5 enzyme X X ZFYVE1 zinc finger, FYVE domain containing 1 other X X ZNF474 zinc finger protein 474 other X ZNF572 zinc finger protein 572 other X X ZSCAN29 zinc finger and SCAN domain containing 29 other X X ZZEF1 zinc finger, ZZ‐type with EF‐hand domain 1 other X X
Supplementary Table 14: Association of radiation‐related DSB hotspots with neoplasia and cancer in NHEJ‐impaired NHDF‐A in p6
Symbol Entrez Gene Name Type(s) Neoplasia Cancer glutamate receptor, G‐protein coupled GRM8 X X metabotropic 8 receptor OR11H12 olfactory receptor, family 11, other X (includes others) subfamily H, member 12 tumor protein p53 binding TP53BP2 other X X protein 2 transcription BCOR BCL6 corepressor X regulator 134 SUPPLEMENT
activated leukocyte cell adhesion ALCAM other X X molecule sodium channel, voltage‐gated, SCN11A ion channel X type XI, alpha subunit heart development protein with HEG1 other X EGF‐like domains 1
Supplementary Table 15: Association of radiation‐related DSB hotspots with neoplasia and cancer in NHEJ‐impaired NHDF‐A in p8
Symbol Entrez Gene Name Type(s) Neoplasia Cancer MSH6 mutS homolog 6 enzyme X X BRCA1 associated RING transcription BARD1 X X domain 1 regulator glial cell derived GDNF growth factor X neurotrophic factor LTA4H leukotriene A4 hydrolase enzyme X TMEM220 transmembrane protein 220 other X chromosome 7 open reading C7orf61 other X frame 61 H3F3C H3 histone, family 3C other X ZNF433 zinc finger protein 433 other X EPHA6 EPH receptor A6 kinase X
Supplementary Table 16: Distribution of therapy‐related DSB hotspots
# #Hotspots #Hotspots Type of DSB # #Hotspots Cell Type Associated associated to associated to Hotspot Hotspots intragenic Genes Cancer Neoplasia Radiation‐ 54 88 34 67 77 NHDF‐A related Doxorubicin‐ 171 389 168 266 314 related Radiation‐ 0 0 0 0 0 related p4 NHDF‐A Radiation‐ 25 40 14 6 4 +NHEJi related p6 Radiation‐ 98 180 56 2 8 related p8
135 SUPPLEMENT
Supplementary Table 17: Average number of histones per DSB hotspot. Nat: Naturally‐occurring DSB hotspot; Overlap: Overlapping DSB hotspot; IR: Radiation‐related DSB hotspot.
NHDF‐A p4 p6 p8 Histone Nat Overlap IR Nat Overlap IR Nat Overlap IR Modification H3K4me1 3.2 6.2 3.2 3.0 8.0 3.9 3.8 5.9 4.8 H3K4me2 2.7 3.8 2.3 3.0 6.6 3.7 2.3 4.3 4.3 H3K4me3 1.9 4.0 4.0 3.0 4.9 3.5 3.0 3.4 3.5 H3K9ac 3.5 3.6 3.3 3.0 7.7 3.7 4.0 5.1 4.7 H3K27ac 3.0 3.8 2.8 3.0 7.6 5.1 2.5 5.4 5.7 H3K36me3 2.6 4.4 5.5 3.0 5.5 4.1 2.4 5.0 4.3 H4K20me1 2.8 4.5 3.3 3.0 7.5 3.6 3.0 5.9 4.2 H3K9me3 3.8 3.5 3.0 3.0 5.8 4.0 3.0 4.7 4.1 H3K27me3 2.7 4.1 1.5 3.0 5.7 3.5 6.0 4.4 4.4 H3K79me2 2.5 5.3 3.2 3.0 7.7 3.8 2.5 5.7 4.3
5.3 Figure Index
5.3.1 Figures
FIGURE 1: DNA DAMAGING AGENTS AND THE DNA DAMAGE RESPONSE IN MAMMALIAN CELLS. 2 FIGURE 2: CLASSICAL AND ALTERNATIVE NHEJ‐REPAIR IN MAMMALIAN CELLS. 6 FIGURE 3: INDUCTION OF IONIZATIONS AND DNA DAMAGE BY HIGH‐ AND LOW LET IRRADIATION. 8 FIGURE 4: OVERVIEW ABOUT REPAIR OF CLUSTERED DNA DAMAGES AND REPAIR OUTCOMES. 9 FIGURE 5: THE TOP2 CATALYTIC CYCLE 10 FIGURE 6: THE RNA AND PROVIRAL GENOME OF HIV‐1. 13 FIGURE 7: LENTIVIRAL LIFE CYCLE. THE LENTIVIRAL LIFE CYCLE CAN BE DIVIDED INTO EARLY AND LATE STAGE. 15 FIGURE 8: PROPOSED MECHANISM OF IDLV‐MEDIATED DSB CAPTURING. 17 FIGURE 9: SCHEMATIC OF 5’ LAM‐PCR FOR THE AMPLIFICATION OF RETROVIRAL INTEGRATION SITES. 35 FIGURE 10: SURESELECET TARGET ENRICHMENT WORKFLOW FOR ILLUMINA SEQUENCING. 40 FIGURE 11: EXPERIMENTAL SCHEME FOR THE INDUCTION, TAGGING AND DETECTION OF RADIATION‐INDUCED DSB SITES IN SURVIVING CELL POPULATIONS. 54 FIGURE 12: KINETICS OF DSB INDUCTION AND REPAIR BY IMMUNOSTAINING OF H2AX FOCI IN (A) A549 AND (B) NHDF‐A AFTER IRRADIATION WITH 1 AND 4GY X‐RAYS. 55 FIGURE 13: RELATIVE CELL VIABILITY OF NHDF‐A (A) AND HELA (B) AFTER SEVEN DAY TREATMENT WITH ETOPOSIDE AND DOXORUBICIN (MTT ASSAY). 56 FIGURE 14: IDLV‐TRAPPING OF RADIATION‐INDUCED DSB SITES IN CANCER CELL LINES AND HUMAN PRIMARY FIBROBLASTS. 57 FIGURE 15: CHEMICAL INHIBITION OF NHEJ‐REPAIR ACTIVITY INCREASES INTEGRATION OF IDLV IN IRRADIATED CELLS. 58 FIGURE 16: FREQUENCY OF IDLV‐TRANSDUCED EGFP+‐CELLS TREATED WITH LD5 AND LD10 DOXORUBICIN AT DAYS 28, 36 AND 42. 59 136 SUPPLEMENT
FIGURE 17: DISTRIBUTION OF VECTOR‐GENOME JUNCTIONS IN IDLV VECTOR. (A) DESIGN AND DISTRIBUTION OF IDLV‐VECTOR BAITS. 61 FIGURE 18: SYNCHRONIZATION OF NHDF‐A IN G1/G0, S AND G2/M PHASE OF THE CELL CYCLE. 63 FIGURE 19: TDT‐MEDIATED POLYU LABELING OF RADIATION‐INDUCED DSB SITES. 64 FIGURE 20: GEL PICTURE OF LAM‐PCR AMPLICONS. 65 FIGURE 21: LAM‐DST FOR IN SITU LABELLING OF DSB SITES. 67 FIGURE 22: LAM‐PCR ON SSDNA LINKER‐LABELLED DSB SITES IN CCR5 ZFN‐TREATED 293T CELLS. 68 FIGURE 23: CHROMOSOMAL DISTRIBUTION OF (A) RADIATION‐ AND (B) DOXORUBICIN‐INDUCED DSB IN NHDF‐A. 69 FIGURE 24: DSB FREQUENCY IN REFSEQ GENES AND AT THE TRANSCRIPTION START SITE (TSS). 70 FIGURE 25: TRANSCRIPTIONAL ACTIVITY DOES NOT INFLUENCE DSB SITE DISTRIBUTION IN IRRADIATED A549 CELLS. 72 FIGURE 26: ENRICHMENT OF RADIATION‐ AND DOXORUBICIN‐INDUCED DSB SITES IN DNASEI HS. 74 FIGURE 27: DISTRIBUTION OF DSB SITES INSIDE DHS IN NHDF‐A. 74 FIGURE 28: DISTRIBUTION OF DSB SITES WITH RESPECT TO CHROMATIN AND HISTONE MODIFICATIONS. 75 FIGURE 29: COMBINATIONS OF HISTONE MODIFICATIONS AT DSB SITES IN IRRADIATED AND DOXORUBICIN‐ TREATED NHDF‐A. 76 FIGURE 30: INGENUITY PATHWAY ANALYSIS (IPA) OF RADIATION‐ AND DOXORUBICIN‐INDUCED DSB SITES. 78 FIGURE 31: TOP TEN DSB SITES IN IRRADIATED NHDF‐A WITH ACTIVE (CONTROL) AND IMPAIRED NHEJ‐REPAIR ACTIVITY IN PASSAGE 4, 6 AND 8. 80 FIGURE 32: HISTONE MODIFICATIONS ASSOCIATED WITH THE TOP TEN DSB SITES IN IRRADIATED NHDF‐A IN PASSAGE 4, 6 AND 8. 81 FIGURE 33: DISTRIBUTION OF DSB HOTSPOTS IN A549 AND NHDF‐A. 83 FIGURE 34: THE GENOMIC DISTRIBUTION OF RADIO‐ AND DOXORUBICIN‐RELATED DSB HOTSPOTS. 84 FIGURE 35: CHROMATIN TOPOLOGY AT RADIATION‐ AND DOXORUBICIN‐RELATED DSB HOTSPOTS IN A549 AND NHDF‐A. 86
5.3.2 Supplementary Figures
SUPPLEMENTARY FIGURE 1: IMMUNOSTAINING OF ΓH2AX FOCI IN IRRADIATED A549 AND NHDF‐A 98 SUPPLEMENTARY FIGURE 2: FREQUENCY OF EGFP+ CELLS TRANSDUCED WITH IDLV DURING CULTIVATION AFTER IRRADIATION. 99 SUPPLEMENTARY FIGURE 3: FREQUENCY OF EGFP+ AND IDLV‐TRANSDUCED NHDF‐A WITH IMPAIRED NHEJ‐ REPAIR ACTIVITY DURING EXPANSION IN CELL CULTURE. 99 SUPPLEMENTARY FIGURE 4: FREQUENCY OF EGFP+, IDLV‐TRANSDUCED HELA CELLS TREATED WITH LD25 AND LD50 DOXORUBICIN AFTER 21 DAYS IN CULTURE. 100 SUPPLEMENTARY FIGURE 5: AVERAGE ΓH2AX FOCI NUMBER IN NHDF‐A TREATED CONTINUSLY WITH DOXORUBICIN AND ETOPOSIDE FOR 43H. 100 SUPPLEMENTARY FIGURE 6: CHROMOSOMAL DISTRIBUTION OF RADIATION‐INDUCED DSB SITES IN A549 (A), PC3 (B), U87 (C) AND NHDF‐A WITH IMPAIRED NHEJ‐REPAIR ACTIVITY (D). 101 SUPPLEMENTARY FIGURE 7: DISTRIBUTION OF RADIATION‐INDUCED DSB SITES IN GENES AND AT THE TRANSCRIPTION START SITE (TSS). 102 SUPPLEMENTARY FIGURE 8: ANALYSIS OF THE GENOME SEQUENCE AT DOXORUBICIN‐INDUCED DSB SITES. 102 137 SUPPLEMENT
SUPPLEMENTARY FIGURE 9: FREQUENCY OF LTR DELETIONS IN INTEGRATED IDLV VECTOR COPIES COMPARED TO AN INTEGRASE‐COMPETENT LENTIVIRUS (ICLV). 103 SUPPLEMENTARY FIGURE 10: THE SIGNAL INTENSITY OF THE DNASEI HS DOES NOT INFLUENCE DISTRIBUTION OF RADIATION‐ AND DOXORUBICIN‐INDUCED DSB SITES. 103 SUPPLEMENTARY FIGURE 11: DSB SITES DO NOT CLUSTER IN PROXIMITY TO DNASEI HS, BUT ARE ENRICHED INSIDE DNASEI HS. 104 SUPPLEMENTARY FIGURE 12: DISTRIBUTION OF RADIATION‐INDUCED DSB SITES IN EU‐ AND HETEROCHROMATIN OF NHEJ‐IMPAIRED NHDF‐A. 104 SUPPLEMENTARY FIGURE 13: HISTONE MODIFICATIONS ASSOCIATED TO RADIATION‐INDUCED DSB SITES IN NHEJ‐IMPAIRED NHDF‐A IN THREE PASSAGES. 105 SUPPLEMENTARY FIGURE 14: RADIATION‐INDUCED AND NATURALLY‐OCCURRING DSB IN NHEJ‐IMPAIRED NHDF‐ A CELLS SHOW ASSOCIATION TO GENOMIC REGIONS MARKED BY H3K27ME3 IN COMBINATION WITH H3K4ME3, H3K9AC, H3K27AC AND H3K36ME3. 106 SUPPLEMENTARY FIGURE 15: RADIATION‐INDUCED AND NATURALLY‐OCCURRING DSB IN A549 CELLS SHOW ASSOCIATION TO GENOMIC REGIONS MARKED BY H3K27ME3 IN COMBINATION WITH H3K4ME3, H3K9AC, H3K27AC AND H3K36ME3. 106 SUPPLEMENTARY FIGURE 16: UPSTREAM REGULATORS OF IDLV‐MARKED GENES IN (A) IRRADIATED AND (B) DOXORUBICIN‐TREATED NHDF‐A 107 SUPPLEMENTARY FIGURE 17: KINETICS AND DISTRIBUTION OF DSB HOTSPOTS WITH RESPECT TO CHROMATIN REGIONS. 108 SUPPLEMENTARY FIGURE 18: NUCLEOTIDE FREQUENCY AT INDICATED POSITIONS DOWNSTREAM OF VECTOR‐ GENOME JUNCTION. 108 SUPPLEMENTARY FIGURE 19: CCR5 GENOMIC LOCUS WITH LIGATED SSGRVU5A DNA LINKER. 109
5.4 Table Index
5.4.1 Tables
TABLE 1: HUMAN DISORDERS ASSOCIATED WITH GENOME MAINTENANCE DEFECTS AND ENHANCED CANCER SUSCEPTIBILITY. 3 TABLE 2: CALCULATED LD50 VALUES FOR DOXORUBICIN AND ETOPOSIDE IN NHDF‐A AND HELA. 56 TABLE 3: QUANTIFICATION OF IDLV INTEGRATION EVENTS. 59 TABLE 4: OVERVIEW ABOUT RESULTS OF HISEQ SEQUENCING OF SURESELECT DNA SAMPLES. 60 TABLE 5: NUMBER OF DSB SITES IDENTIFIED AT 1H, 4H, 8H AND 24H AFTER IRRADIATION 62 TABLE 6: SUMMARY OF SEQUENCING RESULTS FROM LAM‐PCR OF CCR5‐ZFN TREATED AND TDT‐LABELED DSB SAMPLES. 65 TABLE 7: OVERVIEW OF RADIATION‐ AND DOXORUBICIN‐INDUCED DSB IN HOTSPOTS IN EXPANDED CELL POPULATIONS. 82
5.4.2 Supplementary Tables
SUPPLEMENTARY TABLE 1: NUMBER OF UNIQUE DSB IN CANCER CELL LINES AND PRIMARY HUMAN FIBROBLASTS. 110 138 SUPPLEMENT
SUPPLEMENTARY TABLE 2: FREQUENCY OF DSB SITES IN REFSEQ GENES, AT TRANSCRIPTION START SITE AND CPG ISLANDS. 111 SUPPLEMENTARY TABLE 3: FREQUENCY OF LENTIVIRAL ‘GT’ DINUCLEOTIDE AT VECTOR‐GENOME JUNCTIONS. 111 SUPPLEMENTARY TABLE 4: FREQUENCY OF DSB SITES LOCATED INSIDE DNASEI HS. 112 SUPPLEMENTARY TABLE 5: FREQUENCY OF DSB SITES LOCATED IN BORDER REGIONS, EU‐ AND HETEROCHROMATIN 113 SUPPLEMENTARY TABLE 6: FREQUENCY AND ENRICHMENT OF DSB AT CTCF BINDING SITES COMPARED TO RANDOM DSB DATA SET. 114 SUPPLEMENTARY TABLE 7: AVERAGE NUMBER OF HISTONE MODIFICATIONS PER DSB IN A549 AND NHDF‐A 115 SUPPLEMENTARY TABLE 8: UPSTREAM REGULATORS ASSOCIATED TO GENES THAT SHOW INTRAGENIC LABELING OF RADIATION‐INDUCED DSB 116 SUPPLEMENTARY TABLE 9: UPSTREAM REGULATORS ASSOCIATED TO GENES THAT SHOW INTRAGENIC LABELING OF DOXORUBICIN‐INDUCED DSB 118 SUPPLEMENTARY TABLE 10: DISTRIBUTION OF DSB SITES IDENTIFIED IN THE THREE PASSAGES. 120 SUPPLEMENTARY TABLE 11: NUMBER AND POSITION OF THE TOP TEN DSB SITES OVER TIME. 121 SUPPLEMENTARY TABLE 12: ASSOCIATION OF RADIATION‐RELATED DSB HOTSPOTS WITH NEOPLASIA AND CANCER 122 SUPPLEMENTARY TABLE 13: ASSOCIATION OF DOXORUBICIN‐RELATED DSB HOTSPOTS WITH NEOPLASIA AND CANCER 124 SUPPLEMENTARY TABLE 14: ASSOCIATION OF RADIATION‐RELATED DSB HOTSPOTS WITH NEOPLASIA AND CANCER IN NHEJ‐IMPAIRED NHDF‐A IN P6 133 SUPPLEMENTARY TABLE 15: ASSOCIATION OF RADIATION‐RELATED DSB HOTSPOTS WITH NEOPLASIA AND CANCER IN NHEJ‐IMPAIRED NHDF‐A IN P8 134 SUPPLEMENTARY TABLE 16: DISTRIBUTION OF THERAPY‐RELATED DSB HOTSPOTS 134 SUPPLEMENTARY TABLE 17: AVERAGE NUMBER OF HISTONES PER DSB HOTSPOT. 135
139 SUPPLEMENT
5.5 Zusammenfassung
DNA Doppelstrangbrüche (DSB) können sowohl natürlich als auch nach Bestrahlung auftreten und stellen eine große Gefahr für die Stabilität des Genoms und Zellviabilität dar. Wird ein DSB nicht effizient repariert, können Veränderungen im Genom wie z.B. Deletionen und Vervielfältigungen auftreten, die ein Markenzeichen von Krebszellen sind. Ist der DNA Schaden allerdings zu groß, kann die Zelle Apoptose induzieren, was das Ziel der Strahlentherapie zur Krebsbehandlung ist. Derzeitige Methoden zur Bestimmung von DSB basieren vor allem auf der Markierung von DNA Reparaturproteinen wie ATM (Ataxia telangiectasia mutated), 53BP1 (53 binding protein 1) und H2AX (phosphoryliertes Histon H2AX) mittels Fluoreszenz‐markierten Antikörpern. Diese Proteine bilden innerhalb weniger Minuten nach der DSB Entstehung mikroskopisch‐sichtbare Zentren an der DSB Stelle, sogenannte Strahlen‐induzierte Zentren (RIF). Obwohl diese Methode Erkenntnisse über die DSB Entstehung und Reparaturkinetik geliefert hat, ist sie limitiert. So kann keine Aussage über die genaue Position des Strangbruchs getroffen werden, da die RIF auf mehrere Megabasen um die DSB Stelle herum vergrößern. Außerdem verschwinden die RIF innerhalb von 24 Stunden nach Bestrahlung, und ermöglichen somit keine weitere Analyse der reparierten DSB Stellen in überlebenden Zellen. Ebenso erlaubt die Immunfärbung keine Aussage über einen möglichen Zusammenhang zwischen der DSB Stelle, der Reparatur und des Schicksals der Zelle. Daher gibt es kaum Informationen darüber, wie Strahlen‐induzierte DSB im Genom verteilt sind, wie solche Schäden zur Radioresistenz führen sowie die Entwicklung von einer normalen Zelle hin zu einer Tumorzelle und die Krebsentstehung beeinflussen. Deshalb stellen wir die Hypothese auf, dass DSB im Genom nicht zufällig verteilt sind und dass die Analyse der genom‐weiten DSB Verteilung nach Bestrahlung in überlebenden Zellen auf Einzelnukleotidebene unser Verständnis über die Mechanismen, die zur Bestrahlungsresistenz führen, verbessert. Um dieses Ziel zu erreichen, wurden in dieser Arbeit neue Methodenansätze basierend auf der Beobachtung entwickelt, dass Non‐Homologous End Joining (NHEJ) Reparatur zum Einbau von DNA Molekülen ins Genom an DSB Stellen führen kann. Für die DSB Markierung wurden daher Integrase‐defiziente lentivirale Vektoren (IDLV) in Zellen eingebracht, die nach Bestrahlung an transient‐auftretenden DSB Stellen ins Genom stabil eingebaut werden. Diese Integrationsstellen wurden anschließend durch LAM‐PCR und Sequenzierung amplifiziert und identifiziert. Insgesamt wurden somit mehr als 20,000 DSB Stellen im Genom bestrahlter Zellen identifiziert und mit natürlich‐auftretenden und Doxorubicin‐induzierten DSB Stellen verglichen. Die DSB Verteilung wurde mit der Transkriptionsaktivität, verschiedenen Chromatin‐Bereichen und Genfunktionen verglichen. Die Analyse ergab, dass in überlebenden Zellpopulationen das DSB‐Verteilungsmuster nicht zufällig ist. Therapie‐induzierte DSB wurden vermehrt in kleinen genomischen Bereichen, sogenannten Hotspots, die in Grenzregionen von Eu‐ und Heterochromatin und in oder in der Nähe von Onkogenen, Tumorsuppressorgenen und Enhancer Regionen liegen, nachgewiesen. Die Ergebnisse dieser Arbeit zeigen, dass Strahlen‐induzierte Mutationen in überlebenden Zellpopulationen in bevorzugten Bereichen auftreten oder Zellen mit Mutationen in solchen Bereichen selektiert werden. Diese Hotspots können eventuell die Wahrscheinlichkeit für die Entstehung von Therapieresistenzen erhöhen. 140 SUPPLEMENT
5.6 Summary
DNA double strand breaks (DSB) can occur naturally as well as after irradiation and pose a serious threat to genome stability and cell viability. Failure to efficiently repair DSB can result in genomic alterations including deletions and amplifications, which are hallmarks of cancer. If the DNA damage is too severe, apoptosis is induced, which is the main goal in cancer radiotherapy. Current methods for detection of DSB sites following irradiation are mainly based on the immunostaining of DNA repair proteins such ATM (Ataxia telangiectasia mutated), 53BP1 (53 binding protein 1), and H2AX (phosphorylated histone H2AX) by fluorescent‐labeled antibodies. These repair proteins form microscopically‐visible foci at the DSB site referred to as radiation‐ induced foci (RIF) within minutes after DNA damage induction. Even though immunostaining has provided valuable insights into the mechanisms and kinetics of DSB induction and repair, this method has several limitations. Since RIF span up to several megabases around the DSB site, the precise location of the DSB cannot be determined. Moreover, DNA repair proteins disassemble from the DSB site within 24 hours after DSB induction and do not enable DSB site analysis in radiation‐surviving cells. Furthermore, immunostaining does not provide any molecular information on DSB localization, repair outcome and cell fate decision. Hence, up to now, there is little information on how radiation‐induced and repaired DSB sites are distributed, how radiation‐ induced DNA damage is being survived and how these damages induce radiation‐resistance, cell transformation and carcinogenesis. Thus, we hypothesize that the distribution of radiation‐induced DSB sites may not be random and that analyzing the genome‐wide distribution of radiation‐induced DSB in surviving cells at single‐ nucleotide resolution will improve our understanding of the mechanisms of radioresistance. In order to achieve this goal, new methodological approaches were developed based on the observation that DNA repair by non‐ homologous end joining (NHEJ) can lead to the genomic integration of episomal unrelated DNA molecules at the DSB sites. In this thesis, integrase‐deficient lentiviral vectors (IDLV) were introduced into target cells and became stably incorporated into the genome at radiation‐induced DSB sites. The integration sites were subsequently amplified and identified by LAM‐PCR and sequencing. In total, more than 20,000 DSB sites in irradiated and expanded cancer cell lines and primary fibroblasts were obtained and compared to naturally‐occurring and doxorubicin‐induced DSB. Using this DSB data set, the DSB distribution with respect to transcriptional activity, chromatin structures, as well as gene function and networks was analyzed. In radiation‐survivor cells the distribution of DSB was non‐random. DSB sites appeared more frequently in genes with specific functions, in genes regulated by specific regulatory networks and in accessible genomic regions. Radiation‐induced DSB frequently accumulated in genomic hotspots, which were identified at eu‐ and heterochromatin boundaries as well as near oncogenes, tumor‐suppressors and enhancer elements. The results presented in this thesis indicate that radiation‐induced genomic mutations preferably occur or are selected for specific genomic regions in surviving cells. These genomic hotspots of induced DNA damage may increase the likelihood of radioresistance mechanisms and therapy‐induced carcinogenesis. 141 SUPPLEMENT
5.7 Abbreviations
Abbreviation Meaning 3’ 3 prime 5’ 5 prime 53BP1 p53‐binding protein 1 A431 Epidermoid carcinoma cells A549 Adenocarcinomic human alveolar basal epithelial cells ac Acetylated ALD X‐linked adrenoleukodystrophy AP Abasic Sites AT Ataxia telangiectasia ATM Ataxia‐Telangiectasia Mutated ATP Adenosine triphosphate ATR Ataxia‐ and Rad3‐related BER Base Excision Repair Bio Biotinylated BLAST Basic Local Alignment Search Tool BLAT BLAST‐like alignment tool BLS Bloom’s Syndrome bp Base pair BRCA1/2 Breast Cancer 1/2 BSA Bovine Serum Albumin Cas9 CRISPR associated nuclease 9 CAT Chloramphenicol Acetyltransferase CCR5 C‐C chemokine receptor type 5 CD4 Cluster of differentiaton 4 cDNA complementary DNA Chr Chromosome CMV Cytomegalovirus CpG Cytosine and guanine separated by a phosphate CRISPR Clustered Regularly Interspaced Short Palindromic Repeats CtIP CtBP‐interacting protein CXCR4 C‐X‐C chemokine receptor type 4 D64V Aspartic Acid at amino acid position 64 changed to Valine DAPI 4’,6‐diamidino‐2‐phenylindole ddC Dideoxycytidine DDR DNA damage response DHS DnaseI hypersensitive site DKFZ German Cancer Research Center DMEM Dulbecco’s Modified Eagle Medium DNA Dioxyribonucleic acid DNA‐PKcs DNA‐dependent protein kinase, catalytic subunit DKC Dykeratosis congenita dNTP Deoxynucleoside triphosphate 142 SUPPLEMENT
Dox Doxorubicin DPBS Dulbecco’s Phosphate Buffered Saline DPBST Dulbecco’s Phosphate Buffered Saline with 0.2% Tween‐20 DSB DNA double strand break dsDNA double‐stranded DNA dUTP Deoxyuridine triphosphate EC Euchromatin EDTA Ethylenediaminetetraacetic acid eGFP Enhanced green fluorescent protein env Envelope ES Embryonic stem cells Expo Exponential FA Fanconi anemia FACS Fluorescent Activated Cell Sorting FCS Fetal calf serum fwd Forward G1/G0 G1/G0 cell cycle phase G2/M G2/mitosis cell cycle checkpoint g gravitation G segment Gate segment gag Group specific antigen Gp41/120 Glycoprotein 41/120 Gy Gray h Hour H2AX Histone 2A variant X
H2O bidest Double‐distilled water HC Heterochromatin HEK Human Embryonic Kidney cells HeLa Henrietta Lacks HISAP High‐throughput Insertion Site Analysis Pipeline HIV Human Immundeficiency Virus HIT Heidelberg Ion Therapy Center HNSCC Head and neck squamous cell carcinomas HR Homologous Recombination HTLV Human T‐cell lymphotropic virus ICL Interstrand crosslink ICLV Integrase‐Competent Lentivirus IDLV Integrase‐Deficient Lentivirus IgG Immunglobuline G IN Integrase enzyme IPA Ingenuity Pathway Analysis IR Ionizing radiation IMDM Iscove’s Modified Dulbecco’s Medium kb kilobases kDa kilo Dalton 143 SUPPLEMENT
keV kilo electron Volt l Liter LAM‐DST Linker‐Amplification‐Mediated DSB‐Trapping LAM‐PCR Linear Amplification mediated‐PCR LB Luria‐Bertani broth LC Linker cassette LD Lethal Dose LET Linear Energy Transfer LTR Long terminal repeat LV Lentivirus m Milli (10‐3) M Molar Mb Megabase me1/2/3 Mono‐/di‐/tri‐methylated min Minute miRNA MicroRNA ml Milliliter Mlu Micrococcus luteus MMEJ Microhomology‐mediated end joining MMR Mismatch repair Mnase Micrococcal nuclease MOI Multiplicity of infection MPC Magnetic particle concentrator Mre11 Meiotic recombination 11 mRNA Messenger RNA Mse Micrococcus species MTT 3‐(4,5‐dimethylthiazol‐2‐yl)‐2,5‐diphenyl tetrazolium bromide MYO Myoglobin n Nano (10‐9) NBS Nijmegen breakage syndrome NCBI National Center for Biotechnology Information NCT National Center for Tumor Diseases NER Nucleotide Excision Repair ng Nanogram NGS Next generation sequencing NHDF‐A Adult normal human dermal fibroblasts NHEJ (c/a) Classical/alternative Non‐Homologous End Joining nm Nanometer nr Non‐restrictive OH Hydroxyl group P Phosphate residue p Passage PARP1 Poly(ADP‐ribose)‐Polymerase 1 PBS Primer binding sites PC3 Human prostate cancer cell line 144 SUPPLEMENT
PCR Polymerase Chain Reaction PD Pyrimidine dimers PEI Poly‐ethylene Inositol Pen/Strep Penicillin/Streptomycin PFA Paraformaldehyde PGK Phosphoglycerate Kinase 1 promoter PI Propidium Iodide pol DNA polymerase PPT Polypurin tract pUC plasmid University of California q‐RT‐PCR Quantitative real time‐polymerase chain reaction R Redundant RBE Relative Biological Effect RCL Recombination competent lentivirus RECQ Recombination Q RefSeq Reference sequence Rep Replicate rev Reverse RNA Ribonucleic acid RNA‐seq RNA sequencing RNAi RNA interference ROS Reactive Oxygen Species RRE Rev responsive element RSV Rous sarcoma virus rpm rotations per minute RT Room Temperature RTS Rothmund Thomson Syndrome S S cell cycle phase S.O.C. Super Optimal broth with Catabolite repression sec Seconds Ser Serine SIN Self‐inactivating SSB Single‐strand break ssDNA single‐stranded DNA T segment Transport segment TALEN Transcription activator‐like effector nucleases Taq Thermus aquaticus TAR Trans‐activation response element TBE Tris‐borate‐EDTA Tdt Terminal deoxynucleotidyl transferase TE Tris‐EDTA TFBS Transcription factor binding site Thr Threonine TOP1/2 Topoisomerase 1/2 Tsp Thermus species 145 SUPPLEMENT
TSS Transcription start site TUNEL Tdt‐mediated dUTP nick end labeling U3/5 Unique 3’/5’ U87 Human primary glioblastoma cells UC University of California UCSC University of California, Santa Cruz UV Ultraviolet light VSV‐G Vesicular Stomatitis Virus – protein G WPRE Woodchuck hepatitis virus post‐transcriptional regulatory element w/o Without WS (Alternative) Werner Syndrome XLF XRCC4‐like factor XP Xeroderma pigmentosum XRCC4 X‐ray repair cross‐complementing protein 4 ZFN Zinkfinger nuclease α Alpha β Beta Gamma µ Micro (10‐6) µl Microliter µM Micromolar ψ Psi packaging signal
146 SUPPLEMENT
5.8 References
1. Ciccia, A. and S.J. Elledge, The DNA damage response: making it safe to play with knives. Mol Cell, 2010. 40(2): p. 179‐204. 2. Jackson, S.P. and J. Bartek, The DNA‐damage response in human biology and disease. Nature, 2009. 461(7267): p. 1071‐8. 3. Wolters, S. and B. Schumacher, Genome maintenance and transcription integrity in aging and disease. Front Genet, 2013. 4: p. 19. 4. Kumar, R., et al., Chromatin modifications and the DNA damage response to ionizing radiation. Front Oncol, 2012. 2: p. 214. 5. Cleaver, J.E., Defective repair replication of DNA in xeroderma pigmentosum. Nature, 1968. 218(5142): p. 652‐6. 6. Cleaver, J.E., Xeroderma pigmentosum: a human disease in which an initial stage of DNA repair is defective. Proc Natl Acad Sci U S A, 1969. 63(2): p. 428‐35. 7. Madhusudan, S. and D.M. Wilson III, DNA Repair and Cancer: From Bench to Clinic. 2013, Boca Raton, FL, USA: CRC Press Taylor & Francis Group. 718. 8. Huertas, P. and S.P. Jackson, Human CtIP mediates cell cycle control of DNA end resection and double strand break repair. J Biol Chem, 2009. 284(14): p. 9558‐65. 9. Chapman, J.R., M.R. Taylor, and S.J. Boulton, Playing the end game: DNA double‐strand break repair pathway choice. Mol Cell, 2012. 47(4): p. 497‐510. 10. Botuyan, M.V., et al., Structural basis for the methylation state‐specific recognition of histone H4‐K20 by 53BP1 and Crb2 in DNA repair. Cell, 2006. 127(7): p. 1361‐73. 11. Sartori, A.A., et al., Human CtIP promotes DNA end resection. Nature, 2007. 450(7169): p. 509‐14. 12. Yu, X., et al., BRCA1 ubiquitinates its phosphorylation‐dependent binding partner CtIP. Genes Dev, 2006. 20(13): p. 1721‐6. 13. Lukas, J., C. Lukas, and J. Bartek, More than just a focus: The chromatin response to DNA damage and its role in genome integrity maintenance. Nat Cell Biol, 2011. 13(10): p. 1161‐9. 14. Cowell, I.G., et al., gammaH2AX foci form preferentially in euchromatin after ionising‐radiation. PLoS One, 2007. 2(10): p. e1057. 15. Goodarzi, A.A., P. Jeggo, and M. Lobrich, The influence of heterochromatin on DNA double strand break repair: Getting the strong, silent type to relax. DNA Repair (Amst), 2010. 9(12): p. 1273‐82. 16. Jakob, B., et al., DNA double‐strand breaks in heterochromatin elicit fast repair protein recruitment, histone H2AX phosphorylation and relocation to euchromatin. Nucleic Acids Res, 2011. 39(15): p. 6489‐ 99. 17. Reynolds, P., et al., The dynamics of Ku70/80 and DNA‐PKcs at DSBs induced by ionizing radiation is dependent on the complexity of damage. Nucleic Acids Res, 2012. 40(21): p. 10821‐31. 18. Lomax, M.E., L.K. Folkes, and P. O'Neill, Biological consequences of radiation‐induced DNA damage: relevance to radiotherapy. Clin Oncol (R Coll Radiol), 2013. 25(10): p. 578‐85. 19. Hoeijmakers, J.H., DNA damage, aging, and cancer. N Engl J Med, 2009. 361(15): p. 1475‐85. 20. Mimori, T. and J.A. Hardin, Mechanism of interaction between Ku protein and DNA. J Biol Chem, 1986. 261(22): p. 10375‐9. 21. Getts, R.C. and T.D. Stamato, Absence of a Ku‐like DNA end binding activity in the xrs double‐strand DNA repair‐deficient mutant. J Biol Chem, 1994. 269(23): p. 15981‐4. 22. Deriano, L. and D.B. Roth, Modernizing the nonhomologous end‐joining repertoire: alternative and classical NHEJ share the stage. Annu Rev Genet, 2013. 47: p. 433‐55. 23. Shiloh, Y. and Y. Ziv, The ATM protein kinase: regulating the cellular response to genotoxic stress, and more. Nat Rev Mol Cell Biol, 2013. 14(4): p. 197‐210. 24. Margulies, M., et al., Genome sequencing in microfabricated high‐density picolitre reactors. Nature, 2005. 437(7057): p. 376‐80. 25. Bentley, D.R., et al., Accurate whole human genome sequencing using reversible terminator chemistry. Nature, 2008. 456(7218): p. 53‐9. 26. Metzker, M.L., Sequencing technologies ‐ the next generation. Nat Rev Genet, 2010. 11(1): p. 31‐46. 27. Hall, E.J. and A.J. Giaccia, Radiobiology for the Radiobiologist. 5th edition ed. 2000, Philadelphia, PA, USA: Lippincott Williams&Wilki. 28. Denekamp, J., T. Waites, and J.F. Fowler, Predicting realistic RBE values for clinically relevant radiotherapy schedules. Int J Radiat Biol, 1997. 71(6): p. 681‐94. 29. Goodhead, D.T., Initial events in the cellular effects of ionizing radiations: clustered damage in DNA. Int J Radiat Biol, 1994. 65(1): p. 7‐17. 147 SUPPLEMENT
30. Allen, C., et al., Heavy charged particle radiobiology: using enhanced biological effectiveness and improved beam focusing to advance cancer therapy. Mutat Res, 2011. 711(1‐2): p. 150‐7. 31. Asaithamby, A., B. Hu, and D.J. Chen, Unrepaired clustered DNA lesions induce chromosome breakage in human cells. Proc Natl Acad Sci U S A, 2011. 108(20): p. 8293‐8. 32. Urushibara, A., et al., DNA damage induced by the direct effect of He ion particles. Radiat Prot Dosimetry, 2006. 122(1‐4): p. 163‐5. 33. Ward, J.F., Some biochemical consequences of the spatial distribution of ionizing radiation‐produced free radicals. Radiat Res, 1981. 86(2): p. 185‐95. 34. Anderson, J.A., et al., Participation of DNA‐PKcs in DSB repair after exposure to high‐ and low‐LET radiation. Radiat Res, 2010. 174(2): p. 195‐205. 35. Harper, J.V., J.A. Anderson, and P. O'Neill, Radiation induced DNA DSBs: Contribution from stalled replication forks? DNA Repair (Amst), 2010. 9(8): p. 907‐13. 36. Groth, P., et al., Homologous recombination repairs secondary replication induced DNA double‐strand breaks after ionizing radiation. Nucleic Acids Res, 2012. 40(14): p. 6585‐94. 37. Pommier, Y., et al., DNA topoisomerases and their poisoning by anticancer and antibacterial drugs. Chem Biol, 2010. 17(5): p. 421‐33. 38. Vos, S.M., et al., All tangled up: how cells direct, manage and exploit topoisomerase function. Nat Rev Mol Cell Biol, 2011. 12(12): p. 827‐41. 39. Connelly, J.C. and D.R. Leach, Repair of DNA covalently linked to protein. Mol Cell, 2004. 13(3): p. 307‐ 16. 40. Quennet, V., et al., CtIP and MRN promote non‐homologous end‐joining of etoposide‐induced DNA double‐strand breaks in G1. Nucleic Acids Res, 2011. 39(6): p. 2144‐52. 41. Gomez‐Herreros, F., et al., TDP2‐dependent non‐homologous end‐joining protects against topoisomerase II‐induced DNA breaks and genome instability in cells and in vivo. PLoS Genet, 2013. 9(3): p. e1003226. 42. Roy, K., et al., Hypoxia relieves X‐ray‐induced delayed effects in normal human embryo cells. Radiat Res, 2000. 154(6): p. 659‐66. 43. Kim, G.J., G.M. Fiskum, and W.F. Morgan, A role for mitochondrial dysfunction in perpetuating radiation‐ induced genomic instability. Cancer Res, 2006. 66(21): p. 10377‐83. 44. Suzuki, K., Yamauchi, M., Suzuki, M., Oka, Y., Yamashita, S., Involvement of Non‐Homologous End‐Joining in Radiation‐Induced Genomic Instability, in Selected Topics in DNA Repair, C. Chen, Editor. 2011, InTech. 45. Suzuki, K., Multistep nature of X‐ray‐induced neoplastic transformation in mammalian cells: genetic alterations and instability. J Radiat Res, 1997. 38(1): p. 55‐63. 46. Goldberg, Z., Clinical implications of radiation‐induced genomic instability. Oncogene, 2003. 22(45): p. 7011‐7. 47. Suzuki, K., S. Kodama, and M. Watanabe, Role of Ku80‐dependent end‐joining in delayed genomic instability in mammalian cells surviving ionizing radiation. Mutat Res, 2010. 683(1‐2): p. 29‐34. 48. Suzuki, K., et al., Radiation‐induced DNA damage and delayed induced genomic instability. Oncogene, 2003. 22(45): p. 6988‐93. 49. Limoli, C.L., et al., Differential induction of chromosomal instability by DNA strand‐breaking agents. Cancer Res, 1997. 57(18): p. 4048‐56. 50. Degrassi, F., M. Fiore, and F. Palitti, Chromosomal aberrations and genomic instability induced by topoisomerase‐targeted antitumour drugs. Curr Med Chem Anticancer Agents, 2004. 4(4): p. 317‐25. 51. Rous, P., A Sarcoma of the Fowl Transmissible by an Agent Separable from the Tumor Cells. J Exp Med, 1911. 13(4): p. 397‐411. 52. Temin, H.M. and S. Mizutani, RNA‐dependent DNA polymerase in virions of Rous sarcoma virus. Nature, 1970. 226(5252): p. 1211‐3. 53. Baltimore, D., RNA‐dependent DNA polymerase in virions of RNA tumour viruses. Nature, 1970. 226(5252): p. 1209‐11. 54. Varmus, H.E., J.M. Bishop, and P.K. Vogt, Appearance of virus‐specific DNA in mammalian cells following transformation by Rous sarcoma virus. J Mol Biol, 1973. 74(4): p. 613‐26. 55. Gallo, R.C., Growth of human normal and leukemic T cells: T‐cell growth factor (TCGF) and the isolation of a new class of RNA tumor viruses (HTLV). Blood Cells, 1981. 7(2): p. 313‐29. 56. Barre‐Sinoussi, F., et al., Isolation of a T‐lymphotropic retrovirus from a patient at risk for acquired immune deficiency syndrome (AIDS). Science, 1983. 220(4599): p. 868‐71. 57. Coffin, J.M., S.H. Hughes, and H.E. Varmus, The Interactions of Retroviruses and their Hosts, in Retroviruses, J.M. Coffin, S.H. Hughes, and H.E. Varmus, Editors. 1997: Cold Spring Harbor (NY). 148 SUPPLEMENT
58. Sakuma, T., M.A. Barry, and Y. Ikeda, Lentiviral vectors: basic to translational. Biochem J, 2012. 443(3): p. 603‐18. 59. Katz, R.A., et al., The avian retroviral IN protein is both necessary and sufficient for integrative recombination in vitro. Cell, 1990. 63(1): p. 87‐95. 60. Coiras, M., et al., Understanding HIV‐1 latency provides clues for the eradication of long‐term reservoirs. Nat Rev Microbiol, 2009. 7(11): p. 798‐812. 61. Mitchell, R.S., et al., Retroviral DNA integration: ASLV, HIV, and MLV show distinct target site preferences. PLoS Biol, 2004. 2(8): p. E234. 62. Wu, X., et al., Transcription start regions in the human genome are favored targets for MLV integration. Science, 2003. 300(5626): p. 1749‐51. 63. Schroder, A.R., et al., HIV‐1 integration in the human genome favors active genes and local hotspots. Cell, 2002. 110(4): p. 521‐9. 64. Zufferey, R., et al., Self‐inactivating lentivirus vector for safe and efficient in vivo gene delivery. J Virol, 1998. 72(12): p. 9873‐80. 65. Naldini, L., et al., Efficient transfer, integration, and sustained long‐term expression of the transgene in adult rat brains injected with a lentiviral vector. Proc Natl Acad Sci U S A, 1996. 93(21): p. 11382‐8. 66. Leavitt, A.D., et al., Human immunodeficiency virus type 1 integrase mutants retain in vitro integrase activity yet fail to integrate viral DNA efficiently during infection. J Virol, 1996. 70(2): p. 721‐8. 67. Koyama, T., et al., DNA damage enhances integration of HIV‐1 into macrophages by overcoming integrase inhibition. Retrovirology, 2013. 10: p. 21. 68. Mardis, E.R., The impact of next‐generation sequencing technology on genetics. Trends Genet, 2008. 24(3): p. 133‐41. 69. Giard, D.J., et al., In vitro cultivation of human tumors: establishment of cell lines derived from a series of solid tumors. J Natl Cancer Inst, 1973. 51(5): p. 1417‐23. 70. DuBridge, R.B., et al., Analysis of mutation in human cells by using an Epstein‐Barr virus shuttle system. Mol Cell Biol, 1987. 7(1): p. 379‐87. 71. Kaighn, M.E., et al., Establishment and characterization of a human prostatic carcinoma cell line (PC‐3). Invest Urol, 1979. 17(1): p. 16‐23. 72. Beckman, G., et al., G‐6‐Pd and Pgm Phenotypes of 16 Continuous Human Tumor Cell Lines ‐ Evidence against Cross‐Contamination and Contamination by Hela Cells. Human Heredity, 1971. 21(3): p. 238‐&. 73. Mosmann, T., Rapid colorimetric assay for cellular growth and survival: application to proliferation and cytotoxicity assays. J Immunol Methods, 1983. 65(1‐2): p. 55‐63. 74. Schindelin, J., et al., Fiji: an open‐source platform for biological‐image analysis. Nat Methods, 2012. 9(7): p. 676‐82. 75. Dupre, A., et al., A forward chemical genetic screen reveals an inhibitor of the Mre11‐Rad50‐Nbs1 complex. Nat Chem Biol, 2008. 4(2): p. 119‐25. 76. Sharungbam, G.D., et al., Identification of stable endogenous control genes for transcriptional profiling of photon, proton and carbon‐ion irradiated cells. Radiat Oncol, 2012. 7: p. 70. 77. Mullis, K., et al., Specific enzymatic amplification of DNA in vitro: the polymerase chain reaction. Cold Spring Harb Symp Quant Biol, 1986. 51 Pt 1: p. 263‐73. 78. Schmidt, M., et al., High‐resolution insertion‐site analysis by linear amplification‐mediated PCR (LAM‐ PCR). Nat Methods, 2007. 4(12): p. 1051‐7. 79. Schmidt, M., et al., Detection of retroviral integration sites by linear amplification‐mediated PCR and tracking of individual integration clones in different samples. Methods Mol Biol, 2009. 506: p. 363‐72. 80. Clark, J.M., Novel non‐templated nucleotide addition reactions catalyzed by procaryotic and eucaryotic DNA polymerases. Nucleic Acids Res, 1988. 16(20): p. 9677‐86. 81. Kent, W.J., BLAT‐‐the BLAST‐like alignment tool. Genome Res, 2002. 12(4): p. 656‐64. 82. Paruzynski, A., et al., Genome‐wide high‐throughput integrome analyses by nrLAM‐PCR and next‐ generation sequencing. Nat Protoc, 2010. 5(8): p. 1379‐95. 83. Arens, A., et al., Bioinformatic clonality analysis of next‐generation sequencing‐derived viral vector integration sites. Hum Gene Ther Methods, 2012. 23(2): p. 111‐8. 84. Abel, U., et al., Analyzing the number of common integration sites of viral vectors‐‐new methods and computer programs. PLoS One, 2011. 6(10): p. e24247. 85. Gabriel, R., et al., An unbiased genome‐wide analysis of zinc‐finger nuclease specificity. Nat Biotechnol, 2011. 29(9): p. 816‐23. 86. Lin, Y. and A.S. Waldman, Capture of DNA sequences at double‐strand breaks in mammalian chromosomes. Genetics, 2001. 158(4): p. 1665‐74. 149 SUPPLEMENT
87. Nobile, C., J. Nickol, and R.G. Martin, Nucleosome phasing on a DNA fragment from the replication origin of simian virus 40 and rephasing upon cruciform formation of the DNA. Mol Cell Biol, 1986. 6(8): p. 2916‐22. 88. Cartier, N., et al., Hematopoietic stem cell gene therapy with a lentiviral vector in X‐linked adrenoleukodystrophy. Science, 2009. 326(5954): p. 818‐23. 89. Capranico, G., K.W. Kohn, and Y. Pommier, Local sequence requirements for DNA cleavage by mammalian topoisomerase II in the presence of doxorubicin. Nucleic Acids Res, 1990. 18(22): p. 6611‐9. 90. Gross, D.S. and W.T. Garrard, Nuclease Hypersensitive Sites in Chromatin. Annual Review of Biochemistry, 1988. 57: p. 159‐197. 91. Voigt, P., W.W. Tee, and D. Reinberg, A double take on bivalent promoters. Genes Dev, 2013. 27(12): p. 1318‐38. 92. Tessitore, A., et al., MicroRNAs in the DNA Damage/Repair Network and Cancer. Int J Genomics, 2014. 2014: p. 820248. 93. Morris, L.G., et al., Recurrent somatic mutation of FAT1 in multiple human cancers leads to aberrant Wnt activation. Nat Genet, 2013. 45(3): p. 253‐61. 94. Wong, C.C., et al., Inactivating CUX1 mutations promote tumorigenesis. Nat Genet, 2014. 46(1): p. 33‐8. 95. Takita, J., et al., Aberrations of NEGR1 on 1p31 and MYEOV on 11q13 in neuroblastoma. Cancer Sci, 2011. 102(9): p. 1645‐50. 96. Balint, I., et al., Cloning and characterisation of the RBCC728/TRIM36 zinc‐binding protein from the tumor suppressor gene region at chromosome 5q22.3. Gene, 2004. 332: p. 45‐50. 97. Saitou, M., et al., Identification of the TCL6 genes within the breakpoint cluster region on chromosome 14q32 in T‐cell leukemia. Oncogene, 2000. 19(23): p. 2796‐802. 98. Thomas, G., et al., A multistage genome‐wide association study in breast cancer identifies two new risk alleles at 1p11.2 and 14q24.1 (RAD51L1). Nat Genet, 2009. 41(5): p. 579‐84. 99. Vainio, P., et al., High‐throughput transcriptomic and RNAi analysis identifies AIM1, ERGIC1, TMED3 and TPX2 as potential drug targets in prostate cancer. PLoS One, 2012. 7(6): p. e39801. 100. Reed, J.A., et al., SKI pathways inducing progression of human melanoma. Cancer Metastasis Rev, 2005. 24(2): p. 265‐72. 101. Geyik, E., et al., Investigation of the association between ATP2B4 and ATP5B genes with colorectal cancer. Gene, 2014. 540(2): p. 178‐82. 102. Szczepanski, M.J., et al., Triggering of Toll‐like receptor 4 expressed on human head and neck squamous cell carcinoma promotes tumor development and protects the tumor from immune attack. Cancer Res, 2009. 69(7): p. 3105‐13. 103. Dupont, W.D., et al., Protein phosphatase 2A subunit gene haplotypes and proliferative breast disease modify breast cancer risk. Cancer, 2010. 116(1): p. 8‐19. 104. Iorns, E., et al., Identification of CDK10 as an important determinant of resistance to endocrine therapy for breast cancer. Cancer Cell, 2008. 13(2): p. 91‐104. 105. Linja, M.J., et al., Expression of androgen receptor coregulators in prostate cancer. Clin Cancer Res, 2004. 10(3): p. 1032‐40. 106. Bi, X., et al., Drosophila caliban, a nuclear export mediator, can function as a tumor suppressor in human lung cancer cells. Oncogene, 2005. 24(56): p. 8229‐39. 107. Li, Q., et al., A death receptor‐associated anti‐apoptotic protein, BRE, inhibits mitochondrial apoptotic pathway. J Biol Chem, 2004. 279(50): p. 52106‐16. 108. Hurov, K.E., C. Cotta‐Ramusino, and S.J. Elledge, A genetic screen identifies the Triple T complex required for DNA damage signaling and ATM and ATR stability. Genes Dev, 2010. 24(17): p. 1939‐50. 109. Hoefig, K.P., et al., Eri1 degrades the stem‐loop of oligouridylated histone mRNAs to induce replication‐ dependent decay. Nat Struct Mol Biol, 2013. 20(1): p. 73‐81. 110. Meeks‐Wagner, D. and L.H. Hartwell, Normal stoichiometry of histone dimer sets is necessary for high fidelity of mitotic chromosome transmission. Cell, 1986. 44(1): p. 43‐52. 111. Adamson, B., et al., A genome‐wide homologous recombination screen identifies the RNA‐binding protein RBMX as a component of the DNA‐damage response. Nat Cell Biol, 2012. 14(3): p. 318‐28. 112. Tang, W., et al., BIRC6 promotes hepatocellular carcinogenesis: Interaction of BIRC6 with p53 facilitating p53 degradation. Int J Cancer, 2014. 113. Neumann, B., et al., Phenotypic profiling of the human genome by time‐lapse microscopy reveals cell division genes. Nature, 2010. 464(7289): p. 721‐7. 114. Singleton, B.K., et al., The C terminus of Ku80 activates the DNA‐dependent protein kinase catalytic subunit. Mol Cell Biol, 1999. 19(5): p. 3267‐77. 150 SUPPLEMENT
115. Huynh, K.D., et al., BCoR, a novel corepressor involved in BCL‐6 repression. Genes Dev, 2000. 14(14): p. 1810‐23. 116. Banath, J.P., S.H. Macphail, and P.L. Olive, Radiation sensitivity, H2AX phosphorylation, and kinetics of repair of DNA strand breaks in irradiated cervical cancer cell lines. Cancer Res, 2004. 64(19): p. 7144‐9. 117. Neumaier, T., et al., Evidence for formation of DNA repair centers and dose‐response nonlinearity in human cells. Proc Natl Acad Sci U S A, 2012. 109(2): p. 443‐8. 118. Bouquet, F., C. Muller, and B. Salles, The loss of gammaH2AX signal is a marker of DNA double strand breaks repair only at low levels of DNA damage. Cell Cycle, 2006. 5(10): p. 1116‐22. 119. Muslimovic, A., et al., Numerical analysis of etoposide induced DNA breaks. PLoS One, 2009. 4(6): p. e5859. 120. Iacovoni, J.S., et al., High‐resolution profiling of gammaH2AX around DNA double strand breaks in the mammalian genome. EMBO J, 2010. 29(8): p. 1446‐57. 121. Markova, E., N. Schultz, and I.Y. Belyaev, Kinetics and dose‐response of residual 53BP1/gamma‐H2AX foci: co‐localization, relationship with DSB repair and clonogenic survival. Int J Radiat Biol, 2007. 83(5): p. 319‐29. 122. Wanisch, K. and R.J. Yanez‐Munoz, Integration‐deficient lentiviral vectors: a slow coming of age. Mol Ther, 2009. 17(8): p. 1316‐32. 123. Wang, M., et al., PARP‐1 and Ku compete for repair of DNA double strand breaks by distinct NHEJ pathways. Nucleic Acids Res, 2006. 34(21): p. 6170‐82. 124. Neal, J.A., et al., Inhibition of homologous recombination by DNA‐dependent protein kinase requires kinase activity, is titratable, and is modulated by autophosphorylation. Mol Cell Biol, 2011. 31(8): p. 1719‐33. 125. Lieber, M.R., The mechanism of human nonhomologous DNA end joining. J Biol Chem, 2008. 283(1): p. 1‐5. 126. Bozas, A., et al., Genetic analysis of zinc‐finger nuclease‐induced gene targeting in Drosophila. Genetics, 2009. 182(3): p. 641‐51. 127. Brussel, A. and P. Sonigo, Analysis of early human immunodeficiency virus type 1 DNA synthesis by use of a new sensitive assay for quantifying integrated provirus. J Virol, 2003. 77(18): p. 10119‐24. 128. Nightingale, S.J., et al., Transient gene expression by nonintegrating lentiviral vectors. Mol Ther, 2006. 13(6): p. 1121‐32. 129. Gabriel, R., et al., Comprehensive genomic access to vector integration in clinical gene therapy. Nat Med, 2009. 15(12): p. 1431‐6. 130. Cornu, T.I. and T. Cathomen, Targeted genome modifications using integrase‐deficient lentiviral vectors. Mol Ther, 2007. 15(12): p. 2107‐13. 131. Seo, J., et al., Genome‐wide profiles of H2AX and gamma‐H2AX differentiate endogenous and exogenous DNA damage hotspots in human cells. Nucleic Acids Res, 2012. 40(13): p. 5965‐74. 132. Seo, J., et al., Genome‐wide reorganization of histone H2AX toward particular fragile sites on cell activation. Nucleic Acids Res, 2014. 42(2): p. 1016‐25. 133. Barlow, J.H., et al., Identification of early replicating fragile sites that contribute to genome instability. Cell, 2013. 152(3): p. 620‐32. 134. Zhou, V.W., A. Goren, and B.E. Bernstein, Charting histone modifications and the functional organization of mammalian genomes. Nat Rev Genet, 2011. 12(1): p. 7‐18. 135. Peng, J.C. and G.H. Karpen, Heterochromatic genome stability requires regulators of histone H3 K9 methylation. PLoS Genet, 2009. 5(3): p. e1000435. 136. Guelen, L., et al., Domain organization of human chromosomes revealed by mapping of nuclear lamina interactions. Nature, 2008. 453(7197): p. 948‐51. 137. Wen, B., et al., Large histone H3 lysine 9 dimethylated chromatin blocks distinguish differentiated from embryonic stem cells. Nat Genet, 2009. 41(2): p. 246‐50. 138. Li, Z., et al., Histone H4 Lys 20 monomethylation by histone methylase SET8 mediates Wnt target gene activation. Proc Natl Acad Sci U S A, 2011. 108(8): p. 3116‐23. 139. Chapman‐Rothe, N., et al., Chromatin H3K27me3/H3K4me3 histone marks define gene sets in high‐ grade serous ovarian cancer that distinguish malignant, tumour‐sustaining and chemo‐resistant ovarian tumour cells. Oncogene, 2013. 32(38): p. 4586‐92. 140. Hanahan, D. and R.A. Weinberg, Hallmarks of cancer: the next generation. Cell, 2011. 144(5): p. 646‐74. 141. Kraemer, A., et al., MicroRNA‐mediated processes are essential for the cellular radiation response. Radiat Res, 2011. 176(5): p. 575‐86. 142. Kraemer, A., et al., Cell survival following radiation exposure requires miR‐525‐3p mediated suppression of ARRB1 and TXN1. PLoS One, 2013. 8(10): p. e77484. 151 SUPPLEMENT
143. Wang, J., et al., Repression of ATR pathway by miR‐185 enhances radiation‐induced apoptosis and proliferation inhibition. Cell Death Dis, 2013. 4: p. e699. 144. Zhang, J., et al., Loss of microRNA‐143/145 disturbs cellular growth and apoptosis of human epithelial cancers by impairing the MDM2‐p53 feedback loop. Oncogene, 2013. 32(1): p. 61‐9. 145. Skvortsov, S., et al., Rac1 as a potential therapeutic target for chemo‐radioresistant head and neck squamous cell carcinomas (HNSCC). Br J Cancer, 2014. 110(11): p. 2677‐87. 146. Meistrich, M.L., et al., Low levels of chromosomal mutations in germ cells derived from doxorubicin‐ treated stem spermatogonia in the mouse. Cancer Res, 1990. 50(2): p. 370‐4. 147. Goodarzi, A.A., et al., ATM signaling facilitates repair of DNA double‐strand breaks associated with heterochromatin. Mol Cell, 2008. 31(2): p. 167‐77. 148. Olszak, A.M., et al., Heterochromatin boundaries are hotspots for de novo kinetochore formation. Nat Cell Biol, 2011. 13(7): p. 799‐808. 149. Yuen, K.W., B. Montpetit, and P. Hieter, The kinetochore and cancer: what's the connection? Curr Opin Cell Biol, 2005. 17(6): p. 576‐82. 150. Vader, G., et al., Protection of repetitive DNA borders from self‐induced meiotic instability. Nature, 2011. 477(7362): p. 115‐9. 151. Motea, E.A. and A.J. Berdis, Terminal deoxynucleotidyl transferase: the story of a misguided DNA polymerase. Biochim Biophys Acta, 2010. 1804(5): p. 1151‐66. 152. Srinivasan, M., D. Sedmak, and S. Jewell, Effect of fixatives and tissue processing on the content and integrity of nucleic acids. Am J Pathol, 2002. 161(6): p. 1961‐71. 153. Wang, K., et al., Molecular engineering of DNA: molecular beacons. Angew Chem Int Ed Engl, 2009. 48(5): p. 856‐70. 154. Crosetto, N., et al., Nucleotide‐resolution DNA double‐strand break mapping by next‐generation sequencing. Nat Methods, 2013. 10(4): p. 361‐5. 155. Tsai, S.Q., et al., GUIDE‐seq enables genome‐wide profiling of off‐target cleavage by CRISPR‐Cas nucleases. Nat Biotechnol, 2015. 33(2): p. 187‐97. 156. Wang, X., et al., Unbiased detection of off‐target cleavage by CRISPR‐Cas9 and TALENs using integrase‐ defective lentiviral vectors. Nat Biotechnol, 2015. 33(2): p. 175‐8. 157. Frock, R.L., et al., Genome‐wide detection of DNA double‐stranded breaks induced by engineered nucleases. Nat Biotechnol, 2015. 33(2): p. 179‐86.
152 SUPPLEMENT
5.9 Publications and congress attendances
Publications:
Mientus, M., Brady, S., Angelov, A., Zimmermann, P., Wemheuer, B., Schuldes, J., Daniel, R., Liebl, W. (2013) Thermostable Xylanase and β‐Glucanase Derived from the Metagenome of the Avachinsky Crater in Kamchatka (Russia). Current Biotechnology 2(4): 284‐293(10)
Ni, M., Zimmermann, P.K., Kandasamy, K., Lai, W., Li, Y., Leong, M.F., Wan, A.C., Zink, D. (2012) The use of a library of industrial materials to determine the nature of substrate‐dependent performance of primary adherent human cells. Biomaterials 33(2): 353‐364
Zimmermann, P.K., Chiblak, S., Weber, C., Käppel, C., Debus, J., von Kalle, C., Abdollahi, A., Schmidt, M., Nowrouzi A. Genome‐Wide Detection of Radiation‐Induced DNA Double Strand Breaks and Long‐Term Genomic Instability at Single Nucleotide Resolution. (Manuscript submitted)
Nowrouzi, A., Aoi, T., Herbst, F., Deichmann, A., Arens, A., Kara, N., Zimmermann, P.K., Abel, U., Utikal, J., Glimm, H., Schmidt, M., Yamanaka, S., von Kalle, C. Retrovirus Insertion in iPSC Identifies Genes Facilitating Somatic Reprogramming. (Manuscript in preparation)
153 SUPPLEMENT
Conference attendances:
March 1‐6, 2015 Keystone Symposia on Molecular and Cellular Biology: Genomic Instability and DNA Repair
Whistler, British Columbia, Canada
Contribution: Poster
December 19th, 2014 DKFZ PhD Poster Presentation
Heidelberg, Germany
Contribution: Poster
July 9th‐11th, 2014 19th Annual PhD Retreat
Weil der Stadt, Germany
Contribution: Oral Presentation
May 21‐24, 2014 17th Annual Meeting of the American Society of Gene and Cellular Therapy
Washington, D.C., USA
Contribution: Poster
October 25‐28, 2013 European Society of Gene and Cellular Therapy and Stem Cell Clonality and Genome Stability Retreat Collaborative Congress
Madrid, Spain
Contribution: Poster
March 3‐8, 2013 Keystone Symposia on Molecular and Cellular Biology: Genomic Instability and DNA Repair
Banff, Alberta, Canada
Contribution: Poster
154 DANKSAGUNG
6. DANKSAGUNG
Auf diesen letzten Seiten möchte ich die Möglichkeit nutzen, allen meinen Kollegen, Freunden und meiner Familie für die spannenden Jahre meiner Promotion zu danken.
Als erstes möchte ich mich bei Herrn Prof. Dr. Christof von Kalle dafür danken, dass ich diese spannende Arbeit in seiner Abteilung am NCT durchführen durfte. Außerdem möchte ich mich für die konstruktiven Diskussionen sowie für die zahlreichen Möglichkeiten, die Ergebnisse der Arbeit auf zahlreichen internationalen Konferenzen zu präsentieren, bedanken.
Außerdem gilt mein Dank Herrn Prof. Dr. Stefan Wiemann für seine Bereitschaft als Zweitgutachter dieser Doktorarbeit sowie als Mitglied meines Thesis Advisory Commitees zu agieren. Daneben möchte ich auch Herrn Prof. Dr. Jürgen Debus für seine Teilnahme am Thesis Advisory Commitees danken und Herrn Dr. Dr. Amir Abdollahi für die tolle Kooperation und seine Bereitschaft als Teilnehmer meines Prüfungskommitees aufzutreten. Ein großer Dank geht auch an Herrn Dr. Manfred Schmidt für die Betreuung, seine Ratschläge, die Korrekturen und die Möglichkeit diese Arbeit in seiner Arbeitsgruppe durchzuführen.
Eine besondere Stellung bei dieser Arbeit nimmt Dr. Ali Nowrouzi ein, dem ich für die direkte Betreuung im Labor und die Unterstützung in den vergangenen dreieinhalb Jahren in der Abteilung zutiefst dankbar bin und der mir ein sehr guter Freund geworden ist. Ohne die wertvollen Ratschläge, Diskussionen, die gemeinsamen Unternehmungen in der Freizeit und das leckere Essen wäre diese Arbeit nicht möglich gewesen. Unsere gemeinsame Zeit werde ich in sehr guter Erinnerung behalten und vermissen. Vielen Dank für die schöne Zeit.
Vielen Dank möchte ich auch an meine ehemaligen und derzeitigen Kollegen Malaika Rabenstein, Ina Kutschera, Christine Lulay, Steffie Laier, Kai Lukat, Hannah Pfeff, Lena Fuchs, Martina Burk, Steffie Laufs, Eliana Ruggiero, Anna Paruzynski, Simone Scholz, Tim Rath, Cynthia Bartholomä, Richard Gabriel, Irene Gil‐Farina, Stefan Wilkening, Christine Engeland, Friederike Herbst und Annika Mengering für die tolle Atmosphäre, die Unterstützung und die lustigen Stunden im Labor richten. Weiterhin bedanke ich mich bei meinen ehemaligen Masterstudenten Paranchai Boonsawat und Mahdi Akbarpour, die einen großen Beitrag geleistet haben, sowie den Abteilungen von Amir Abdollahi, Stefan Fröhling/ Claudia Scholl und Hanno Glimm. Ein spezielles Dankeschön geht auch an Raffaele Fronza und Saira Afzal, die mich bei der Auswertung der bioinformatischen Daten unterstützen.
Zuletzt geht ein besonderer Dank an meine Freunde und meine Familie, die mich immer in meinem Studium und während meiner Promotion unterstützt haben. Eine herausragende Stellung nehmen meine Eltern und mein Bruder ein. Trotz der Entfernung und der wenigen Treffen im Jahr habt ihr mich während des Studiums und der Doktorarbeit immer unterstützt und mir immer einen wunderbaren Aufenthalt zu Hause in Berlin ermöglicht. Ohne eure Hilfe wäre ich nicht soweit gekommen. Dafür gebührt euch ein besonderer Dank! Außerdem danke ich Elke, Melanie, Peter und Fabi für die tollen und erholsamen Stunden in der Pfalz, das leckere Essen und Trinken sowie die Möglichkeit im Weinberg mitzuhelfen. Eine Reise zu euch ist immer ein schöner Urlaub! Zudem geht ein großer Dank an meine Frau Sarah, mit der ich die gemeinsame Zeit im Studium, in der Freizeit und während Promotion genießen durfte. Du warst und bist mir immer eine große Unterstützung, und ohne dich wäre mir das Studium, die schöne Zeit in Singapur und die Promotion nicht möglich gewesen. Ich freue mich sehr auf unsere weiteren Herausforderungen und eine tolle Zukunft.
Ein großer Dank geht auch an meine Freunde. Vor allem zu nennen ist dabei Christian Weber, für die Unterstützung in der Zellkultur, die zahlreichen lustigen Pausen, die MTB Trips, Superbowl Abende und die viele 155 DANKSAGUNG
Zeit, die wir sonst noch zusammen hatten! Danke möchte ich auch an Friederike Knipping, Christine Käppel, Christian Großardt, Aaron und Jessica Tideman richten, mit denen ich viele schöne gemeinsame Stunden in vielen Kletterhallen, Kinos, Bars und Restaurants in der Rhein‐Neckar Region verbringen durfte. Ein großer Dank geht auch an meine Berliner Freunde Sylvia und Simone sowie Gerrit und Daria. Außerdem möchte ich meiner Hockeymannschaft für die vielen Trainingseinheiten und Reisen sowie Erol für die unzähligen Trainingsstunden auf dem Spinningrad, an der Langhantel und dem TRX danken, die mir halfen, die Gedanken wieder neu zu sortieren, neue Energie zu tanken und endlich richtig fit zu werden.
Natürlich darf man bei einer Arbeit über die Verteilung von DNA Schäden nicht die zahlreichen Putzkräfte vergessen, die jeden Morgen meine Konzentration am Computer mit dem Staubsauger störten. Euch widme ich dieses Werk. Ohne Kaffee und Döner wäre diese Arbeit auch nicht durchführbar gewesen, daher danke ich allen Barista in Heidelberg und der Dönerbude am Nahkauf. Außerdem bin ich den Polizistinnen und Polizisten in Heidelberg dankbar, dass sie jeden Morgen meine rote Autoplakette in der Umweltzone übersehen haben und mir so viel Strafgeld erspart blieb. Danke auch an die Förster im Odenwald für die guten Mountainbike Bedingungen auf den zahllosen Trails rund um den Weißen Stein und den Königsstuhl. Ohne eure herausragende Arbeit wäre so manch ein Ride nicht ganz so flowig geworden. Als großer Sportjunky muss ich natürlich auch einen Dank an die Sportartikelhersteller Nike, Asics und Adidas sowie den Hockeymarken TK und Voodoo richten, die mich super unterstützten viel Zeit im Labor zu sparen. Zu tiefstem Dank verpflichtet bin ich auch der „Voice of G100“ Jury für die durchweg positive Kritik an meinem Gesangstalent. Sie haben mich dazu inspiriert meine Gesangskarriere auch weiterhin mit größtem Einsatz in anderen Laboratorien dieser Welt zu verfolgen. Ich werde euch auch weiterhin zahlreiche Kostproben meiner Gesangseinlagen kostenlos zukommen lassen. Zuletzt bedanke ich mich bei den Programmdirektoren von RTL, RTL2 und Vox, die mich mit ihrem allabendlichen Programm daran erinnern, dass es definitiv die richtige Entscheidung war zu studieren und promovieren.